Online Responsibility: Bad Samaritanism and the Influence of Internet Mediation

Abstract

In 2008 a young man committed suicide while his webcam was running. 1,500 people apparently watched as the young man lay dying: when people finally made an effort to call the police, it was too late. This closely resembles the case of Kitty Genovese in 1964, where 39 neighbours supposedly watched an attacker assault and did not call until it was too late. This paper examines the role of internet mediation in cases where people may or may not have been good Samaritans and what their responsibilities were. The method is an intuitive one: intuitions on the various potentially morally relevant differences when it comes to responsibility between offline and online situations are examined. The number of onlookers, their physical nearness and their anonymity have no moral relevance when it comes to holding them responsible. Their perceived reality of the situation and ability to act do have an effect on whether we can hold people responsible, but this doesn’t seem to be unique to internet mediation. However the way in which those factors are intrinsically connected to internet mediation does seem to have a diminishing effect on responsibility in online situations.

Introduction

On November 19, 2008 a 19 year old boy from Florida, Abraham Biggs, committed suicide in front of his webcam by taking an overdose of prescription medication. The webcam kept running while he lay on his bed, dying. He had alerted people on a forum he frequented of his intentions, and had posted a link to where the webcam could be viewed.

There was significant online discussion, both on the forum and in the live chat feature with the webcam feed. There was a lot of joking about the situation and there were also people who egged him on. After a long time, people began to worry as they noticed he no longer seemed to be moving. Eventually, a forum member from India managed to alert the police, and not long after that the police officers could be seen storming into the room. By then Abraham had been dead for many hours.

The incident aroused public outrage. Apparently 1,500 people had been watching that webcam at some point, and part of the outrage was that none of them interfered until it was too late.¹ The media presented the case of Abraham Biggs as a modern version of the case of Kitty Genovese, whose fatal stabbing in 1964 was passively observed by many witnesses thus arousing great public revulsion.

Kitty’s death prompted several states of the United States to create legislation for what is dubbed bad samaritanism (cf. Feinberg 1984; McIntyre 1994): the failure to aid in an emergency situation can, in some cases, render people legally liable. We hold bad Samaritans responsible because they should have taken action and they are blameworthy for not having done so. It is the responsibility of the onlookers that I will focus on in this paper.

Is the Abraham Biggs case indeed a contemporary version of Kitty Genovese’s case? Apart from some obvious differences between the ways in which these two tragic deaths took place, what makes Biggs’ case seem special is the role played by the internet. This paper will investigate the moral significance of internet as a mediator of events and whether it diminishes or adds to our responsibility for them. One’s intuitions go both ways: on the one hand, it would have been very easy for the average onlooker to call the police; suggesting surely they were responsible. On the other hand, can we really assign responsibility to a person in India for preventing the suicide of someone living in Florida, US?

Online experiences are often contrasted with “real life” experiences. This indicates that internet mediated interaction is apparently conceived as something that is “not real”.² Of course, when things are conceived of as “not real”, then what is morally permissible changes too. For instance, shooting down a digital villain in a first person shooter game is (arguably) morally no different from taking a pawn in a game of chess. However, in reality the internet is becoming more and more a part of our daily lives, not only of our private lives but of our professional lives as well. If it is true that internet mediation has a negative effect on our responsibilities, in the sense that it makes us less responsible, this might have consequences for legislation as well. Furthermore, if it is clear why and how the internet has an effect on our moral behaviour, we might be able to design interfaces in a way that enables us to take responsibility online in the same way we can take responsibility offline.³

The case of Abraham Biggs is the starting point from where I will make comparisons with the case of Kitty Genovese and see if there are differences that are possibly relevant for the degree to which an agent is morally responsible that are specific to internet mediation. It is clear that there are some morally relevant differences between the cases that are not due to internet mediation. For instance, Abraham committed suicide and can arguably be said to have wanted to die, whereas it is pretty certain that Kitty did not seek her own death. I will leave aside the question of whether one has a duty to interfere with someone who wants to commit suicide and instead I will assume for the sake of the argument that one does have such a duty to intervene.

A very compelling method of constituting morally relevant differences is provided by Peter Unger, who tweaks example cases in such a way that there is only one variable that is different in each. He then tests intuitions about these cases (Unger 1996). There are a number of difficulties with this approach (Philips 2007), the first of which is the difficulty of how to identify which factors are morally relevant; second is the problem of then constructing cases in such a way that this factor is the only difference between the cases; and third, figuring out how to ensure that the intuitions about those cases can be trusted, that is to say, that the constructed cases are such that it is likely that one would have moral intuitions about them.⁴ Unger, by using intuitions and showing how they can go awry, then argues against the notion that intuitive responses to specific cases can determine what is right and wrong in those cases. But as Singer (1999) points out, the claims about the unreliability of intuitions is something that should be empirically investigated much better before such strong conclusions can be drawn.

To help determine whether it is the internet that makes a morally relevant difference, I will present a hypothetical contemporary Kitty Genovese case, where the situation is placed in an internet mediated setting. After explaining how the onlookers can be said to have been responsible, I will discuss the various candidate factors for moral differences and compare the cases in that respect. Note that it is the moral intuitions about the onlookers that I will compare and not the actual responses of the onlookers. The method of comparison is an intuitive one. Although it would have been ideal to construct a set of examples in the way that Peter Unger did in order to probe our intuitions about whether online and offline situations make a difference, it turned out that it is very difficult to construct examples in such a way that there is only one aspect differing between them. Instead I will work with the examples that I have, accepting that there are multiple difference between the cases, and try my best to provide a rationale for those intuitions. It is for this reason that I give a comprehensive account of the case of Abraham Biggs: different intuitions about both this case and the online situated Kitty case can help determine whether it is indeed the internet mediation that is making the difference, rather than other factors that are not exclusive to online situations. To the extent that moral philosophy is dependent on empirical cases, it is important to have a precise description of these empirical cases: if the description does not correspond with what happened, the insights gained from it do not inform us about the reality of such a situation. The cases will be presented next.

Abraham Biggs

In order to understand the circumstances on of Abraham Biggs’ death some background information on some features of the internet is needed. I will provide this information and then I will give an account of what happened on the night of his demise. In doing so, I will probably provide more information than is strictly necessary for my purposes, but to avoid misrepresenting the case as has happened with the case of Kitty Genovese, I want to be as accurate as possible.

The Bodybuilding Forum

The forum Abraham Briggs subscribed to and announced his intentions on, bodybuilding.com, is part of a commercial site which sells bodybuilding products and also provides information about the subject. It is a widely popular site: they currently claim to have approximately 250,000 unique visitors every day. The part of the site that hosts the forum provides users with the opportunity to discuss issues with other members. The forum can be read by everyone who visits the site, but in order to respond to topics, one has to become a member. This is a simple process that does not include any verification of the identity of the user. People generally enter a pseudonym with which they will be identified on the site. There are currently over two million members of the forum. Those are not all active; that is, they do not all post regularly (if at all). Those who are active make up the community of the forum. The common language is English, but members come from wherever there is internet access.

The forum is subdivided into categories that deal with bodybuilding issues. People can reply to topics started by other people or start their own topics (which are referred to as threads). In principle one can start any topic under any category, but there are two mechanisms by which things are kept orderly: first, there are moderators, who have the ability to edit and delete posts. Moderators are volunteers who have been granted these privileges by the administrator of the forum. Second, the community will react negatively to posts that are put up in inappropriate sections, which ensures people are careful about where they start their topics.

Maintaining one’s “reputation” within the community is another important aspect of the forum. Members of the forum can assign positive and negative reputation points to other members, the weight of which depends on the number of posts, reputation points already owned, and years of membership. Having a certain number of positive reputation points unlocks areas of the site that are not accessible to those who don’t have enough reputation points (“reps”). The reputation point system provides an incentive for members to post helpful posts, to avoid posting comments that will get negative “reps” and to be prolific in posting, which keeps the forum alive.

Although forums are generally dedicated to a particular subject, there usually is a section dedicated to off-topic discussion. The header for these discussions is generally “Miscellaneous” which is often abbreviated to “misc”. Often, the “misc” section is the one in which there is the most activity (i.e. posts). To illustrate, on the bodybuilding forum, of the approximately 51 million posts that are currently posted, close to 25 million posts are in the “misc” section under “more general categories”. The next most popular sections are “supplements” under “main forums” with almost seven million posts and “teen bodybuilding” under “Specifically for you” with just over five million posts.⁵

The discussion in the “misc” section can be about anything, and since members feel a sense of community with other regularly posting members, private things often get discussed here, with subjects varying from trouble with girlfriends (although there are many female members, the majority of the participants are male and estimated to be around 20), to school, parents, and other personal problems. The “misc” section is therefore also the section that attracts the most “trolls”, as they are called. Those are people who intentionally try to disrupt communities, mainly by baiting and riling people. Posts about private matters make those members vulnerable for ridicule, which is what trolls seek out. The tone of the discussion in the “misc” section varies a lot, often with sympathetic and encouraging messages from other members, but there are also often instances of ridiculing and taunting. The general atmosphere seems to be one in which people are expected to know that both kinds of responses are possible and they are expected to accept it.

Posts in forums remain there until removed, which generally does not happen. Replies to a thread can be posted long after a thread is started (unless the thread has been closed for comments). Although the postings on a “hot” topic can be very fast and exceed a post a second, the medium is not really suited for fast communication, mainly because the thread loses coherence with such postings: during the time someone is typing a response in a thread, others have already responded, and it becomes difficult to follow. The medium is also unsuitable for live streaming of webcam information. Not only does it require a different technology to post a live webcast than a video file, the transient nature of webcasts does not fit with the more permanent nature of a forum post.⁶

Live Webcasting

There are various sites on the internet that enable people to broadcast live images from a webcam (live webcasting) on the net. The site that Abraham used was Justin.tv. With this site, one needs to register in order to be able to webcast from one’s webcam. In order to view the webcam, however, no registration is necessary. Amongst the information and facilities that the site offers to viewers are the webcam feed, the number of viewers and the total number of views since the webcast started. Next to the webcam images is a box where viewers can chat in real time. This means that people can type messages in a box under the chat window, which will then instantly appear in the chat window, along with the messages that other people have typed. This is a more immediate form of communicating than typing messages on a forum, and it is also more transient. Generally, the chat screen just shows a maximum number of lines, and everything that was typed before becomes inaccessible (except for the administrator of the site, and then only if the software is set up to maintain archives). At the time Abraham Biggs webcast his suicide, it was not necessary to register in order to be able to chat in the chat window, although it was possible to register, in which case people’s chosen screen name would appear before the line posted in the chat window.

Reconstruction of the Case

Having explained the nature of the media, I shall now proceed to reconstruct the circumstances of Abraham Biggs’ death. This is somewhat difficult for reasons that will become apparent later.

Abraham Biggs was a 19 year old college student in Florida and he was an active member in the “misc” section on the bodybuilding forum.⁷ Abraham was diagnosed with bipolar disorder and had posted about committing suicide before. For instance, in December 2007 he started a thread about how he felt down about himself and how he had attempted suicide before. The responses were not particularly nasty or disparaging. On the whole, Abraham did pretty well with “reps”: he got second rank with the number of reputation points he obtained.

At 2:35 AM⁸ on November 19, 2008, Abraham, who was known as “CandyJunkie” on the forum, started a new thread “I’m trippin on bars on my webcam”, where he posted a link to his webcam images on Justin.tv. He added a message, saying “thoughts of suicide are there but no worrys I’m not gonna kill myself”. A few minutes later he specified that he had taken a “roxie” which is short for the drug roxicodone, a narcotic pain reliever. At 2:55 AM he posted “[…] I’m gonna take the last bar I have and a bunch of roxies, and some lexapro which I don’t think will do anything, I’ll be dead within a few hours.” The substances Abraham referred to are prescription medicines: “bar” is the urban term for Xanax tablets (alprazolam), a benzodiazepine which is prescribed for anxiety disorders; Lexapro is the brand name for escitalopram oxalate which is a Selective Serotonine Reuptake Inhibitor (SSRI) type of antidepressant. From later accounts on the forum, it appears that in this thread a lot of goading took place, along with expressions of concern. It seems that Abraham also posted a suicide note on the forum. There appear to be no records of what happened exactly on camera at that moment, but presumably Abraham took the pills and then lay down on his bed with his back to the camera.

Meanwhile, a lively discussion took place in the chat window beside the camera image on Justin.tv, and from later accounts on the forum, the tone was rather jocular. There was discussion about whether he took enough pills to kill himself, and also whether something should be done, but the latter appeared to be less than serious. From later threads on the forum, it appears that it was initially possible to see Abraham breathing but that at a certain point, people were not sure whether he was breathing anymore.

One 17 year old forum member from India—which has a time difference of 10.5 h from Florida in winter—came home late from work and caught up with what had been going on in the “misc” and came across the threads about “CandyJunkie” and got alarmed. It is on the discussion in this thread on the forum after the events and a blog post by “Bulker” that most of this reconstruction is based.⁹ “Bulker” also clicked the link to Justin.tv and determined that what he saw was not fabricated and realised that he couldn’t see Abraham breathing. He then “[…] scanned the Internet and found out his name, number, where he lived (not exact address though) and pics” (“Bulker” 2008).¹⁰ This information he posted on the forum, and half an hour later he also supplied the telephone number of the Miami police. Since people still insisted that Abraham was faking it, “Bulker” decided to take action himself. First he sent an e-mail to the Miami police, but that e-mail bounced, so he stole his father’s cell phone and called the Miami police. He spoke to four people, none of whom grasped the gravity of the situation. They gave him a different phone number to call, but then the international credit of his father’s mobile account ran out, so he stole his mother’s cell phone and called that number, where he was informed that three people had called about the issue in the meantime.

Fifteen minutes later, the Miami police arrived at Abraham’s room—at approximately 3 PM in Florida and 1:30 AM in India. They first threw an object at the lifeless body and when Abraham’s body didn’t respond, they could be seen to enter the room and move in front of the camera. They checked the body and turned around and put something over the lens of the camera. It was then that people chatting alongside the screen on Justin.tv realised this was serious and the “lolz”¹¹ were replaced by dismay. What happened next is that the people on the forum who had been egging Abraham on rapidly started to delete their posts. Justin.tv also became aware of the matter and deleted both the images that were broadcast and the transcript of the chat screen that had been running along with it, claiming by way of justification that the video had violated their policies. Those two events have hampered the reconstruction of what had actually happened.

The next day, brief accounts of the circumstances surrounding Abraham’s death appeared in the media, and public outrage ensued, this reaction being similar to that following the reporting of the circumstances of the death of Kitty Genovese, which I will describe next.

Kitty Genovese

With over a thousand publications inspired by it, the case of Kitty Genovese has become the showcase of social psychology when it comes to diffusion of responsibility and bystander apathy. The case as it was originally presented inspires disgust at the callousness and apathy of onlookers. In order to have a more accurate comparison with the case of Abraham Biggs, I will set the case in an online situation: the online case will be an intermediary between Kitty’s original case and the case of Abraham. First I shall sketch the original case.

The Original Case

The murder of Kitty Genovese in the early hours of March 13, 1964 was described in the New York Times 2 weeks after it happened (Gansberg 1964). “For more than half an hour 38 respectable, law-abiding citizens in Queens watched a killer stalk and stab a woman in three separate attacks in Kew Gardens”, the article begins, and it continues to depict the case. Kitty got home from her job as a bar manager at approximately 3:20 AM. She parked her car and started to walk the 30 m towards her apartment door, but then she saw a man coming towards her and she started to move away. The man overtook her and stabbed her in the back twice at which Kitty cried out that she had been stabbed. Several of the neighbours’ lights came on and one opened his window and shouted at the attacker to “let the girl alone”, at which the attacker moved away. Kitty, badly wounded, was making her way to the back entrance of her apartment building when the attacker came back and stabbed her again, at which she cried out again. Lights came on from neighbouring windows and a window was opened and the attacker got in his car and left at 3:35 AM. Kitty then made it into the building and collapsed at the foot of the stairs. The attacker returned and found her there and stabbed her again.

The police received a phone call from a neighbour at 3.50 AM and they arrived at the scene in 2 min, but that was too late to save Kitty. It turned out that 39 people had witnessed some part of the assault and only one of them took action by calling the police, albeit too late.

People later said that they didn’t want to get involved or that they thought it was just a lovers’ quarrel.

However, De May and Manning et al. point out that the account in the New York Times is rather inaccurate and misleading (cf. De May 2004; Manning et al. 2007). At most 12 of the witnesses looked out of their windows and saw something taking place, the other witnesses just admitted to having heard cries. Considering there was a bar nearby which often was the source of noisy brawls in the middle of the night, it is not that surprising that people did not give it more attention. Furthermore, there were not three but two attacks, one in front of the building and one at the back of the building, the latter taking place inside and hence out of view.

Despite the above, for the sake of this paper I will make an online adaptation of the case as it has been described in the New York Times and as it has been assumed to be in countless publications.

Kitty on Webcam

Suppose that we place the case of Kitty Genovese on the internet, in such a way that the only witnesses to the event are the people who happen to be watching the webcam. There are many webcams that are aimed at a particular street or park publicly viewed on websites, and it is not unthinkable that an attack like that on Kitty Genovese could be thus witnessed. In order for the people witnessing the attack to be the only ones to witness it, it is more feasible to let the attack take place in a park, rather than in a street in a residential area (because then people might see it from their window, as happened in the original case). So the case is as follows: on a website there is a webcam view of a certain path in a specified park in New York and it is approximately 3 AM in the morning. What can be seen is a man apparently assaulting a woman with an object that looks like a knife. Something scares the attacker away; this cannot be due to some response from webcam viewer, as it is a one way view. The woman staggers on and a few minutes later the attacker returns.

Most street view web cams do not have a microphone, so sound is not conveyed. For our hypothetical online case I suppose the sound is not so relevant, since contrary to the offline situation, people would be clicking on a link to watch a site on purpose, so their attention would not need to be drawn by sounds.

Contrary to broadcasts on Justin.tv, street view cameras generally do not have a counter of concurrent onlookers. This will fit in with our online situated Genovese case, because in the original situation people would not have a clear idea of how many people were witnessing the scene either, although they were probably aware that some were, because one of them opened his window and yelled at the attacker.

In our hypothetical online case, the woman dies on the spot and one of the onlookers does inform the police, but as with the original case, this is too late.

As with the original case, it later transpires that at the time of the crime, 39 online viewers were watching the events take place.

Although I already pointed out some obvious differences from the original case, there are two other differences that are worth noting. First, contrary to the original case, the onlookers are not necessarily located near the scene of the crime. In fact, they might not only be in different countries, but even in different time zones. This has practical relevance in the sense that it might not be easy to find the emergency number of the police at the scene of the crime.¹² The second important difference is that in the online case, it is much easier to substantiate one’s worry that there might be a crime taking place by simply taking snapshots of the screen. If people were reluctant to call because they were not certain of what they saw was a crime and did not want to raise a false alarm, this might lower the threshold for actually making the call.

Now that the cases are spelled out, it is time to look at how the onlookers might have been responsible for the untimely death of a fellow human being.

Responsibility of the Onlookers

In order to figure out whether the internet makes a difference when it comes to responsibility, it is necessary to make at least a rudimentary sketch of how the onlookers might be responsible and might be held responsible. This outline will provide a useful basis for establishing which factors may make a morally relevant difference in online situations compared to offline situations. There are many ways in which people can be responsible and can be held responsible and part of the reason for this plurality is that responsibility is a complex concept, or even a complex of concepts (Vincent 2011).

The moral outrage after the murder on Kitty Genovese did not focus on the grisly details of the stabbing, but on the fact that 38 people apparently sat and watched the mortal stabbing for half an hour and did not take action. We feel that they ought to have done something to intervene, ergo; we feel they had the duty to intervene and thus they had the responsibility to intervene. This kind of responsibility can be assigned by either looking back or looking forward. When we take stock of what has happened after the fact, we can determine whether the subjects either directly caused the harm or were involved in the harmful outcome. When looking forward (that is, considering that the harm in progress is happening right now, rather than in the past), people might be in a position where their role made them responsible for taking action and they are likely liable for not doing what they ought to have done (Vincent 2011).

Looking back on the cases of both Abraham and Kitty, the issue with the onlookers is not that they directly caused the deaths of Abraham or Kitty, but that they omitted doing things that could have prevented further harm and in that sense, they contributed to the outcome of Abraham and Kitty dying.

Can we hold them responsible for that? Apart from having to have the capacity to make decisions, the main condition under which a person can be held responsible is that the person is free from coercion; in other words, that his action or disposition is voluntary. In order for the voluntariness condition to be satisfied, a person must first have control over the situation (the control condition): the action or disposition must originate from within the agent and he must be free either to undertake or not to undertake the action or to have the disposition.¹³ Obviously, not only must people have the freedom to act, but they also must have the ability to act. Second, the person must be aware of what he is doing or bringing about. This is often referred to as the epistemic condition. It must be noted that an attempted rescue is only required if the action does not pose unreasonable risk of harm to the agent: you are not required to sacrifice your life for another.¹⁴

Not only can people be held responsible for actions and dispositions, it is also agreed that people can be held responsible for omissions; if the condition is satisfied that the action that was omitted could have been performed by the person we hold responsible (cf. Fischer and Ravizza 1991).

With regards to omissions, failing to prevent (further) harm can be said to be the cause of that harm if it is true that but for the omission, the harm would not have taken place. If it is true that we have a duty to prevent ourselves from causing harm, then it may also be true that we are not permitted to omit an action if we have the ability to take that action (Feinberg 1987).

With these requirements in mind, we can see how the onlookers in the case of Kitty Genovese can be held responsible. Although some of the onlookers may have misjudged the gravity of the situation, there were at least several who knew that something bad was going on. The action that was required in this situation was to call the police, an action not likely to be harmful to the onlooker. Also, it was a very simple action: all they had to do was dial “0” for operator and ask for the police. It would have been a different situation if their only way of intervening would have been to physically come between Kitty and her attacker, because that would certainly have been dangerous to them.

Now that we have a clearer view on how the onlookers in the case of Kitty Genovese can be held responsible, I will investigate whether the same can be said of onlookers in online situations. Internet mediation may have a direct effect on the measure of responsibility people have but the effects may also be more subtle. For instance, what if the propensity to be irresponsible or to refrain from acting responsibly is similar to a genetic predisposition to alcohol dependency? As long as you are not in a context where you are confronted with alcoholic beverages, you will not run the risk of becoming an alcoholic. However, if you are in a situation where you are likely to consume alcoholic beverages then, due to your genetic disposition, you are quite likely to become an alcoholic. Suppose that exposure to responsibility undermining situations works in a similar vein, then it might well be that in an offline situation, you would be a responsible person, but due to the exposure to situations that undermine your sense of responsibility online, you are not when online.¹⁵ I believe this is something that needs to be empirically researched before we can make any meaningful philosophical deductions.

Factors that are candidates for being of influence on responsibility are discussed in the next section.

Differences Between the Online and Offline Situations

The aim of this section is to identify factors that are of moral relevance when the online and offline situations are compared. The investigation will be restricted to a number of salient differences that are likely to be related to the internet mediation. It is likely that other factors can be identified, but if the factors discussed here alone can be demonstrated to be morally relevant and related to online responsibility, then it can be concluded that internet mediation does indeed affect responsibility.

The five candidates discussed are the number of onlookers, the perceived reality of the situation, the physical distance, the ability of the onlookers to act and anonymity. After individual appraisal of those factors, I will look at how those factors are connected, since it will become clear that it is difficult to look at them in isolation.

The Number of Onlookers

When one is the sole observer of harm in progress and has the ability to do something about it without unreasonable expense to life and limb, then we intuitively feel that we are morally obliged to do what we can in order to avoid further harm. Things get complicated when there are more people who all seem equally capable of doing something, because as the number of onlookers increases, the chances that someone will actually take action decreases significantly (Lataneé and Rodin 1969; Zimbardo 1969), a phenomenon that is now known as the bystander effect.

Although there is a significantly reduced chance of people taking relevant action in a hazardous situation when they are part of a crowd, that does not seem to morally excuse them for not doing something when they could have taken action. It might even be argued that the onlooker is actually more responsible with every additional onlooker, simply because the chances of someone taking action decrease with the increase of the number of onlookers. Then again, this somehow feels unjust to the unsuspecting witness to a harmful situation: it is not his fault that the situation he inadvertently witnesses¹⁶ happens to have multiple witnesses and it seems rather unfair that this could land him with more responsibility. On the other hand, if an agent is the sole witness to a harmful event, that agent has full responsibility for at least considering taking action (including the assessment of what can be done without unreasonable harm to that agent), which may also seem unfair because the agent being a witness is fortuitous: he might equally well not have become a witness. There is an inherent unfairness in stumbling upon such a situation: when it happens, responsibility is instantly thrust upon you. In this light it seems that there is no moral relevance to the number of onlookers in a case in which there was an option to take action without unreasonable risk: whether there is only one or there are a thousand, each individual onlooker has a moral obligation to take action when they can.

Yet intuitively, it appears that people are more offended by cases of preventable harm in which there are many passive onlookers than where there are a few. This is probably explained by the fact that with the increased number of onlookers, there is also an increased probability of the ability of one of those onlookers to intervene. Take for instance Singer’s famous example of the child who is drowning in a pond (Singer 1972). If you, as a witness, are unable to swim, the child will drown.¹⁷ With every additional witness, the chance that someone is able to swim increases, and then the intuition is that therefore, the chance of the child being rescued is increased. But rather than thinking that no one was able to swim in a large crowd that is watching a child drown, it is assumed that many actually could swim but did not use that ability to rescue the child. This makes all the difference in how we feel about it. Considering that people had a duty to help in the first place, this argument does not make a moral difference, though; it just introduces the possibility that some of the onlookers might not have been able to help, increasing the contrast with those who could have and should have helped; again this has nothing to do with their numbers per se.

Is there, then, a morally relevant difference in the number of onlookers online and offline? There does seem to be a difference in how easy or likely it is to achieve a large number of onlookers online, which may have a certain moral relevance.

On the internet it appears to be much easier to gather a crowd. It involves no physical movement of the onlooker; they remain in the same place, from wherever they are accessing the internet. That makes it much easier to accidentally stumble upon places you would not normally physically go (e.g. it is easy to inadvertently land on pornography websites, whereas I wouldn’t be prone to visit a pornography shop in the flesh). It is indeed as simple as just clicking a link that you get presented with somewhere.

In the case of Abraham Biggs, there were 1,500 onlookers¹⁸ and in the original case of Kitty Genovese there were 39. One of the reasons that people became onlookers in the case of Abraham Biggs was that there was an active thread in the forum with the link to the images on Justin.tv, which people clicked on. It might also have been the case that the images of Biggs attracted so many viewers (probably via the forum) that the channel ended up on the top of the page in the “most popular right now” category on Justin.tv, which might have enticed random visitors to click the link to his broadcast.

As easy as people can get to a site on the internet, though, it is equally easy to move away from it again. So people who clicked the link might have been bored with the scene within seconds or less (because apart from the start and end of the broadcast, all that could be seen was a man lying on a bed with his back to the camera). So even though it is easier to gather a crowd online, the crowd is also much more easily dispersed.

Now suppose that we take the online Genovese case and suppose that someone witnessing the attack had posted the link on a popular forum with a relevant comment about something alarming seeming to be happening there. That might well have raised the number of the spectators spectacularly, but also, when people clicked and viewed the moment in between the attacks, they might very well have clicked away again and not have seen anything, yet they still would have been listed as a viewer in the stretch of time of the attack. This is however merely relevant to the type of onlookers: on the internet there is a category of onlooker that can be counted as such (they were on the particular site at the time), but who are totally oblivious as to even having been an onlooker as such, because they never realised that there was something going on that made them responsible to act and who had left before that fact could permeate. I think it is quite possible that this category exists in offline situations too, but due to the fact that it is generally more cumbersome to move oneself physically away from a scene than to click away from a webpage, this might just give an agent the chance to realise that he does have a duty to either figure out more about the situation or to take action.¹⁹ It thus seems not unlikely that the proportion of this type of onlookers is a lot smaller in the offline situation than in the online situation.

It may thus turn out that the number of onlookers is not morally relevant, but there still is an aspect of the number of onlookers that might be so, and that is that the perception of being one of a large number of onlookers. To perceive that you are part of a large crowd might have a psychological effect that changes behaviour in such a way that it becomes morally relevant.

As mentioned before, it is clear that when you are the only observer, the responsibility to take action, or at least contemplate the kind of action needed to intervene, is all yours. So if there is a difference in the situation with the people—on the one hand, in the offline situation in Kew Gardens seeing Kitty being stabbed or, on the other hand, online on Justin.tv seeing Abraham becoming comatose—the case might be upheld that the internet makes a difference. In both cases, though, the onlookers were aware that there were other people witnessing what was going on. The online Genovese case does not help here, because there—although people could not know that others were watching—it obviously is not a feature of the internet per se, since in Abraham’s case, the people were aware that others were seeing what was happening. All in all, there appears to be no morally relevant difference regarding the internet when it comes to the number of spectators, whether the spectators knew there were other spectators or not.²⁰ There does appear to be a difference in the kind of spectators, but it is undetermined yet whether that has moral relevance. What determines the kind of spectator has nothing to do with the number of them, though: it has more to do with how an agent perceives the situation. This concept will be discussed in the next section.

The Perceived Reality of the Situation

The fact that people speak of “real life” as opposed to what they are doing on the internet indicates an important feature of how online reality is perceived, namely that there appears to be something less real about it. In order to elucidate this, let us look at the original Kitty Genovese case again. When the people in Kew Gardens heard screaming outside, there was no doubt that there was someone on the street who was screaming. They knew where the scream was coming from and they could look out the window and see a woman on the street apparently being attacked. Of course, Kitty and her attacker could have been accomplices who acted as if he was attacking her, maybe for some psychology experiment to see whether and how quickly people would try to intervene or maybe even to lure some unsuspecting neighbour out in order to mug and rob them.²¹ This was however an unlikely scenario. The woman might have screamed without being in actual danger, like one neighbour surmised that “it was a lovers’ quarrel”.

Compared to the online Kitty Genovese case, there is a different kind of knowing where the attack took place. The people watching the cam were aware, or could easily be aware that the attack took place in a particular park, because the website would state that the views they are broadcasting are from a camera that is located in that particular park. They could also see that there appeared to be a person being attacked. There is a different level of certainty about the situation, though, that is somewhat related to the distinction of knowledge by acquaintance and knowledge by reference. The mediated way in which the images on the internet are perceived could be said to be not “first hand” experience, but more as having come to you through a third party; as if you were informed about the situation, but not acquainted with the situation through your own senses and from your own perspective. There is much discussion about how certain one can be about observations (cf. Stanley 2008), and about how far you can be certain of sense data at all. For the purpose of this paper I will just stay with the intuitive notion that we are more certain about the observations we make with our own senses, without technological mediation.²²

Something we can be quite sure of is that it is much easier to fake images on the internet than it is in the offline three dimensional world. What is presented as live images could very well be a recording, and manipulation of the presented images is easier on the internet too. This kind of deception is a lot harder in an offline situation: compare for instance “photoshopping”²³ to sleight of hand tricks; the latter takes a lot more skill and practice, whereas “photoshopping” is made very easy by the software.

The two elements of uncertainty regarding images on the internet are that (a) it is not clear that what you see is the same as what you would see when you were there (i.e. they might be manipulated in various ways) and (b) it is not clear that the images you see are happening in real time. This makes it harder to assign responsibility, for if you cannot be certain something is happening, surely it cannot be the case that you are required to do something about it, or at least to consider doing something about it: it does not satisfy the requirements of the epistemic condition which was mentioned earlier.

In the case of Abraham Biggs, there was quite some discussion about whether the images were fake or whether the images were on a loop (that is, one stretch of film that is edited in such a way so that it plays over and over though it seems it is ongoing). However, when considering the online version of the Kitty Genovese case, the likelihood of the images being fake or tampered with is considerably smaller. Websites hosting camera broadcasts of a stretch of scenery generally have the motive to enable people to see that scenery, it is a feature to draw visitors to the website or to keep them there a bit longer. It does not seem very likely that the owner of such a website would set up a prank to deceive people into believing they are witnessing harm in progress.²⁴ With Abraham, that was not inconceivable, because in the context of the “misc” section of the forum, pranks are not uncommon. The video feed from Abraham’s cam and the one in the online Kitty case would differ in that in the former, nothing much appeared to be happening. That might have added to the uncertainty about what it was that was going on.

When presented with uncertainty about a situation, are there ways to disambiguate what is taking place? People are confronted with ambiguous situations all the time and certainly there are heuristics with which we generally make sense of what is happening around us. Among the first things we appeal to is perceived probability: if an event is unlikely to occur, we will tend not to take what is occurring at face value. Take for instance the crowd waiting for the emperor and his new clothes. When the emperor with his invisible new clothes appeared, there was a long silence until one child pointed out that the emperor was naked and then everyone started to laugh. The emperor had no clothes on for the entire time, but the notion that the emperor would appear before the crowd naked is so incredible that it would take a while to sink in.²⁵ This example points to a second way in which we attempt to make sense of unclear situations: we look to others to see what their reaction is. It is a strategy of social confirmation. The people staring at the naked emperor would have looked at each other to see whether others perceived something out of order. In case of an emergency this might even be stronger. Suppose for instance that at a crowded market there was a loud noise and shouting. When you yourself cannot make out what is happening and there are people hurrying back from where the noise originated with alarmed faces, you will interpret the situation as dangerous and will want to move away from the commotion too. On the other hand, if you just see people look and shrug and walk on as normal, you are not likely to be alarmed.

In the offline Kitty case, it was difficult for the spectators to see what other people were doing, since people were in their houses.²⁶ In a similar vein, the people who witnessed the assault on the webcam in the online Kitty case had no way to see each other. However, in the online case, the witnesses could have alerted other online users and told them to go and watch the cam and say what they thought about it. Even though the crime took place at three in the morning, that does not necessarily mean that it was also three in the morning where the viewer was located. Furthermore, it is quite possible that there are contacts online if the person is in the habit of being online at that time. The difference between online contacts and the neighbours in Kew Garden is that you generally are acquainted rather well with your online contacts, and you are aware that they are awake (because they are online). In order to have a chat with the neighbours to clarify the situation, it would mean physically disturbing a person you are not acquainted with in the middle of the night, without being sure he is even awake. In the online case it is therefore probably easier to apply this social confirmation strategy.

In the Biggs case, where people were in a situation where they could interact—using the chat feature next to the camera images,—this strategy went completely awry, because the general consensus appeared to be that the situation was not alarming.

But is this really different from offline situations? There are common examples of such cases, for instance when a homeless person is seen lying unmoving on the street, and everyone just steps around him and pays no attention. It is easy to conclude that the homeless person is just inebriated and will be fine. This might be very wrong, but that says more about the inappropriateness of the strategy than about the role of the internet.

Physical Distance

Another candidate for morally relevant difference is proximity. Can it be that it was of moral relevance that people were physically near?

The view that the people on Justin.tv had of the room of Abraham Biggs gave them no information about where in the world that room was. It is one of the interesting features of the internet that geographical distance becomes quite irrelevant: one can as easily and as instantly talk to people next door as to people on the other end of the world. Although people online generally tend to be curious with regard to the whereabouts of the people they interact with, this interaction does not take place anywhere in particular. Even though technically, there is a physical location where the database of, for instance, the bodybuilding forum is stored, the forum in itself is not anywhere in the world in particular. Anyone on the forum could be anywhere in the world where there is internet access. This curious state of “nowhereness” is interesting with regard to findings about people’s willingness to help victims; because the distance one is removed from a person in need makes a difference. For instance, when Milgram (1963) conducted his shocking obedience experiments,²⁷ one of the independent variables he used was the distance of the subject to the suffering learner. It turned out that distance had a significant effect on the willingness of subjects to administer electrical shocks and he surmised that “distance, time and physical barriers neutralize the moral sense” (Milgram 1974, p. 157). That is to say, the further people were physically removed from the suffering person, the more likely it was that they would continue administering shocks of increasing voltage.

Singer (1972) and Unger (1996) both argue that distance is not morally relevant, but their examples typically demonstrate that apparently, people do have a completely different attitude to suffering in front of their eyes compared to suffering in a remote place.²⁸ In that sense, they seem to confirm what Milgram indicated, that proximity to the subject is of influence on our attitude; in addition they argue that there is something morally wrong with this attitude of letting proximity be of relevance to our judgment.

If we look back at our two real cases and focus on the two people that took action, that is: the neighbour that called the police in Kitty’s case, and the boy in India in Abraham’s case, maybe we can learn something from our intuitions about how much we think they are to be praised for what they did.²⁹ After all, the neighbour was physically near and the boy in India definitely was not. I think that few people would say that the neighbour was actually to be praised for finally calling the police. It is merely that he took the action that all other onlookers failed to take and in that he stands out from the rest. In that sense he is somewhat excused from our blaming of the onlookers, although he was still too late in reacting. The case of the boy in India is quite different; indeed he can be regarded as quite heroic. Where everyone else was apathetic, he carefully observed the situation, deemed it alarming and made a considerable effort to obtain the relevant data (namely, finding out what Abraham’s location was, which was more than the neighbour had to do) and report to the police (which was more complicated than what the neighbour had to do, which was just dial “0” for operator). It was a commendable thing to do, and it is a pity it was too late. I do not think I am the only one who takes this view. Is it true, then, that the distance makes the difference: that is to say, did the neighbour, due to being near, have a greater responsibility than the boy in India? That question is not easy to answer, because there are other factors that are relevant in the case of the Indian boy; for instance, the gravity of the situation was less salient in his case and the fact that it was not so easy to do something about it.³⁰ So what if we take our online Kitty case and surmise that the person taking action was an adult person in India? Unless we construct a bit of an unlikely case—with a person who used to live near the park and knows what number to call and who recently moved to India and happened to venture on the site where the images of the webcam are broadcast at the moment that our hypothetical Kitty was being stabbed—it is not so easy to eliminate other aspects that differ. Since the online Kitty case is more obvious when it comes to the gravity of the situation, compared to the case of Abraham, what we are left with on the one hand, is that it is just more difficult to access the emergency numbers of the different country (because the local number is still different from the number one would have to call from abroad).³¹ On the other hand, I assume it is harder to convince the officer on the phone that what you are seeing is an emergency, since you are nowhere near the event but viewing things on a webcam (as the boy in India discovered). However, if this is what the aspect of distance might be reduced to, it is not the distance per se that is relevant: it is the perceived reality of the situation—which I already covered—and the ability to act, which I will cover next.

The Ability to Act

If a certain situation called for a certain action but you, as an agent, were in no way able to take that action, it seems fair to say that you cannot be responsible for not doing so.³² If we walk by the pond and see the child drowning but we cannot swim, there is no obligation for us to jump in the pond and try to save the child. We ought to look at the alternatives though. If we can attract the attention of someone else or call the police, we may have a responsibility to do that: in being confronted with the state of affairs, we obtain the responsibility to consider the options and do what we can to ameliorate the situation for the drowning child. If there is no one nearby, and you do not carry a phone, then you cannot be responsible for not calling or drawing the attention of another person. If on the other hand, you do carry a phone, but decide that you do not want to be involved and you toss the phone into the pond, then indeed: you cannot make the call, but you are still responsible. The point there is that you deliberately put yourself into a position where you were unable to do what was needed.³³

In the case of Kitty Genovese, the onlooking neighbours, when considering their options may have concluded they were indeed not required to come between Kitty and her attacker, since that would put them in considerable danger. That did not exclude them from the duty to consider what other options they had though, and the most obvious thing to do was to call the police, which would have been very simple (they had phones³⁴ and they only had to dial “0”). In this sense they were fully responsible for their omission to call the police.

In Abraham’s case it was more complicated, because it was not clear where he was exactly; his profile only stated that he was in Florida. There are good reasons not to post your address details on the internet for everyone to find and Abraham had not done so. It could very well be that the onlookers in Abraham’s case, who realised the gravity of the situation and thought that something ought to be done, perceived that they were in fact powerless to do anything, since to find Abraham’s room would surely be like finding a needle in a hay stack and it would be pointless to burden the police with that.

This is different in the online Kitty case, because onlookers there were aware of the exact location of the attack. Their difficulty would be to find the correct number to make the phone call and then to see to it that they effectively communicated to the police that there was an emergency going on. It must be noted that in none of the cases it was certain that calling the police would have prevented further harm: the intervention of the police when they would have appeared to the scene might have been too late. Perhaps then, it was not as salient for everyone that what they ought to do was to call the police, but this is true for both online and offline cases. Considering this, it seems reasonable to state that in general it is harder to figure out exactly where a person in peril is when observed on the net compared to finding that out when you see the person with your own eyes or rather, not mediated by a webcam.³⁵

One thing that needs to be explored in this context, though, concerns the possibility that people can be responsible despite not being able to take action, due to their deliberate incapacitation of themselves. There are various ways in which people may incapacitate themselves, but that does not always necessarily render them responsible for what they made themselves incapable of. Consider the child in the pond again, and consider you can swim in normal circumstances, but you are on your way home from celebrating with friends in the pub and you are inebriated to such extent that you are unable to swim and therefore are unable to save the child. I think most people would agree that this is just tough luck; it would be unreasonable to have to stay in an optimal condition all the time just in the odd case that someone should need your help. One could argue that that is in fact a very good reason to ban alcoholic beverages, but it would also mean having to be fit and alert at all times and then it becomes unreasonable; people do get tired and need sleep. The situation changes when your particular role demands of you that you refrain from incapacitating yourself. For instance, if you are the minder of the child, it is obviously wrong that you got inebriated and failed in your responsibility of taking care of the child. In the same vein, an officer on watch can be held responsible for falling asleep whereas in itself, there is nothing morally wrong about falling asleep (cf. example “Caught Off Guard” in Sher 2009).

It is sheer coincidence that the man who eventually called the police to alert them to the situation of Kitty Genovese appears to have been heavily drinking that night. Maybe if he had not been drinking, he would have been better able to assess the situation and maybe Kitty would have survived the attack. However, considering that he was in the privacy of his own home and did not have a role of taking care of anyone, I doubt we can reproach him for having been inebriated at that particular time; it was not likely that he would have been called upon to take responsibility for anything.

Considering that very often, people who dwell on the internet in the recreational setting of, for instance, the “misc” section of the body building forum, are in the privacy of their own homes, it would seem fair to say that therefore, they are free to incapacitate themselves when they are not in a particular role where they are required to have full control over their faculties. Of course, this is only addressing the issue where they would be incapacitated due to personal circumstances rather than circumstances in general that are related to the internet. There may however, be another case of deliberately incapacitating oneself by engaging in a community where it is, seemingly due to the nature of the medium, impossible to be a responsible person. That is to say, by engaging in a platform where you know you are unable to interfere in the case of an emergency, you may still be responsible, because you will have voluntarily incapacitated yourself. This would, however, be similar to the case of the inebriated passerby to the pond with the drowning child, since in neither case the subject would have a role in which they ought to make sure they did not incapacitate themselves.

All in all, the internet does seem to have a general hampering effect on the ability of people to take action in case of an emergency, and this is partly due to the fact that often essential information, such as the location of a person in peril online, cannot be pinpointed due to the fact that they are there under a pseudonym. This brings me to the next section, where I investigate the role of anonymity.³⁶

Anonymity

It is generally recommended that one be careful about supplying information on the internet that could identify you, for reasons of, for instance, the risk of identity theft and—particularly for children—the risk of being targeted by online predators. A less ominous reason to not be identifiable online is to protect one’s privacy. Forums generally can be read by everyone online and it is not always desirable that utterances in one context spill over in another; for instance, you might discuss a health issue you have on a health forum but you would not want your employer to read your posts; not because you want to deceive your employer, but simply because the information is private.³⁷

It is not surprising, then, that Abraham was on the body building forum under a pseudonym, as are most other users. This anonymity provides for lower risk of the abovementioned dangers, but it also has an effect on how people interact with each other. Zimbardo (1969) defined the effect of anonymity on behaviour as deindividuation: when anonymous,³⁸ people tend to be less self aware, engage less in self-evaluation and are less concerned about social comparison and evaluation. This theory has been revised for computer mediated communication in a Social Identity model of Deindividuation Effects (Christopherson 2007), the discussion of which is outside the scope of this paper. The key point is that anonymity does affect behaviour in positive and negative ways; that is, it affects what actions people take and which actions they refrain from taking. When they are anonymous in computer mediated communication, such as in chat rooms and forums, people generally are less inhibited concerning the use of expletives and strong language. People may also experience a dissociation of their online pseudonyms from their offline selves which makes them feel less accountable for inappropriate online social behaviour (Suler 2004). A combination of these effects may be seen in relation to Abraham’s suicide: the callous egging on and the apathy with regard to trying to intervene. However, this situation is not unique to online situations: the phenomenon of a crowd baiting a person who is about to commit suicide by jumping off a building is rare yet recurring (Mann 1981).

When we compare the online and offline Kitty cases, the onlookers’ measure of anonymity stands out. The neighbours in Kew Garden may not have known who the woman was who was being attacked, but since the location was known and they had a view of that location, it was not very difficult to identify who might have witnessed the event. In fact, this was what the journalist who reported the case in the New York Times did: those neighbours were then put in the uncomfortable position of having to answer why they had not called the police. In contrast, the onlookers in the online case felt themselves completely unobserved and unidentifiable: there was no indication they were being registered (there was no counter of concurrent viewers) and they had not been required to log in. If not for the log of the website, they could have left the scene without a trace. This sense of invisibility would have made it seem easy to get away with not intervening. Given the lack of action in the real case where the chances of being detected would have been perceived as much larger, the result might have been no one taking any action at all. Thus anonymity does affect moral behaviour, but it does not render people less responsible. This is similar to how the number of onlookers reducing the chances of someone helping does not make the individual onlookers less responsible for acting: although there is a change in moral behaviour due to this factor, there is no moral justification for this change in behaviour.

Given that Zimbardo developed a theory concerning people shirking responsibility through anonymity in the late 1960s, the phenomenon is clearly not uniquely related to the internet. It is, however, not only a lot easier to be anonymous on the internet, it is essentially recommended.

The Combination of the Factors

When observed independently, it seems that only the perceived reality of the situation and the ability to act have moral relevance when it comes to the responsibility of the onlookers.

Two things can be said about the combination of the factors, though. First, there is the possibility that factors that may not seem significant may become synergistically significant when grouped together. It may be possible that although in isolation the effects on our responsibility would not reach a threshold where we would say they are significant, taken together they will. If that is true, then internet mediation, due to its inherent accumulation of these factors, makes us less responsible.

Second, the factors are not as distinct as they seemed at first sight: when examining them, it turned out that they are very closely connected with each other. For instance, the number of onlookers may be helpful in determining the reality of the situation, but it might also be detrimental to it due to a failure in the strategy of social confirmation. Awareness of being in a crowd fosters a sense of anonymity of the onlooker. Physical nearness could very well only be relevant due to the fact that you have a better opportunity to confirm the reality of the situation and have better opportunities for taking a successful action. Anonymity affects both the perceived reality of the situation and the ability to act. The ability to act, in turn, depends heavily on your perception of the reality of the situation and also on your perception of what your options are, which is influenced by your proximity to the situation.

Given the interconnectedness of all those factors, then, it is perhaps misleading to consider these factors in isolation. For instance, even if it is true that the only moral relevance of physical distance lies in the side effects of affecting the perception of the reality of the situation and the ability to act, that does not mean that we can claim that the distance does not matter, because distance will always affect perception and ability to act.

As we have seen, taken in isolation, none of the factors discussed are unique to internet mediation. It is well possible that the combination of those factors is not unique either. However, the situations in which you may encounter those factors in concert certainly seem more pervasive on the internet than they are in offline situations: internet mediation is exemplary for those factors.

Conclusion

Does internet mediation make a difference to the extent to which we have a responsibility to take action in a potentially harmful situation? In order to answer that question I contrasted the cases of Kitty Genovese and Abraham Biggs. I examined various salient potentially morally relevant differences between offline and online emergency situations: the number of onlookers, the perceived reality of the situation, the physical distance, the ability to act and anonymity. With each of these factors, there are two germane subsequent questions that relate to the moral relevance of the factor to online responsibility: (a) whether the factor indeed has moral relevance when it comes to responsibility and (b) whether the factor is unique to internet mediation. The number of onlookers, the physical nearness and anonymity of the onlookers in and of themselves seem to have no moral relevance when it comes to the responsibility of those onlookers to take action: they do not directly affect any of the circumstances in which we can hold people responsible. The perceived reality of the situation and the ability to act do affect those circumstances: in order to take responsibility, you have to be aware that the situation is such that it requires you to take action and you have to be able to take action.

However, none of those factors are unique to internet mediation. The reason that there still seem grounds to conclude that internet mediation does influence moral responsibility of onlookers lies in the combination of those factors. It is inherent to internet mediation that one is subjected to all those factors when online and although none of these factors is particularly novel, the fact that they are so inextricably connected to internet mediation makes a difference. Note that this assessment of online responsibility is relative rather than absolute: there appears to be a lessening effect in online situations as compared to offline situations.

For this reason, future studies should investigate what can be done about designing interfaces in such a way that the same level of responsibility we experience offline can be maintained in the course of online interactions as well.