If it is understood that in order to trust something, or for something to be trustworthy, this something must not only have mental statesFootnote 3 but must have mental states in which the interests of the trusting person are represented, then the notion of trust in technological devices cannot make sense (supposing that no technological device is currently a candidate for having mental states ascribed to it, a situation that may change in the future). However, technology is not just a collection of devices. People are involved in technology essentially, and in various ways. First of all, technical devices are designed, manufactured, sold, maintained and, if necessary, repaired by people. This creates a difference between using a bridge to cross a river and using a tree that happened to fall across the river to do so. In this section we discuss whether this allows us to speak of trust in (forms of) technology, if not straightforwardly then perhaps in a derived sense, and whether this move can ultimately be recommended.
Suppose that someone—let’s call him Floris—buys a new kind of coffee machine or some other consumer product and plugs it into a socket. Typically, Floris is confident that it will not blow up in his face. In deciding to use it, he relies on the device doing what the documentation says it will do and not something else that will harm him. The status of something as a technical artifact is particularly associated with the reason for one’s reliance on it. Suppose Floris uses some object, be it natural or artificial, for an incidental purpose—say something heavy for hammering a pole into the ground or for temporarily supporting some other thing. In this case his reliance that things will go the way he intends them to go, assuming his reliance satisfies minimal requirements of rationality, is based on his confirmed judgment that the object has the material properties that enable it to serve his purpose in the way he intends. The crucial word is ‘confirmed’: Floris can use the object because he has checked for himself that the thing has the required properties, possibly by indirectly inferring the presence of these properties from other properties he knows the thing to have, or possibly by seeing someone else use the object in the same way. In contrast, when Floris buys a new kind of coffee machine and plugs it in, he has often no idea what material properties it must have in order to do what it is supposed to do, let alone that he has checked that the product has these properties. Floris’s confidence concerning the product’s expected operation must therefore be grounded in some attitude of his concerning the people who made the device or sold it to him. It is difficult to see what other ground his confidence could have.
But is this confidence an attitude of trust? If his newly bought coffee machine does blow up when plugged in, certainly he is not merely disappointed (supposing he survives)! He will feel this should not have happened, even if it was not anybody’s intention it should. Floris cannot, of course, expect the designer/manufacturer to have mental states concerning his interests or feelings in particular. But they can be expected to know that whoever buys one of their coffee machines, that person is not likely to appreciate it exploding. If it is sufficient for someone to be trustworthy that this person has an interest in building a reputation and will therefore take one’s interests into account, then Floris may trust the designer/manufacturer of his coffee machine to put a product on the market that has an extremely low risk of spontaneous explosion. Suppose that this exhausted Floris’s reasons. Then, it would depend on the sheer number of products they sell, on the costs of avoiding the misbehavior of a product, on the costs to their reputation when a product does misbehave, and on the likelihood of such an occurrence actually damaging their reputation (which would depend again on the quality of the media, of the legal system, and so forth) how much effort would be spent on taking care that Floris's machine should not blow up when plugged in. And when it blows up all the same, Floris may be reminded of the fact that the producing company has made such calculations. He is likely to be disappointed or frustrated. But would he go so far as to say that his trust has been betrayed? Not all disappointments about behavior are betrayals of trust. For a betrayal of trust, something more must be the case than a mere expectation that someone will do X because that person is taking your interests into consideration since it is rational for that person to take your interests into consideration. The something more is typically that this person has let you know, promised you even (at least implicitly), that he or she will act in accordance with your interests.Footnote 4 Betrayal is, in typical cases, nothing else but the breaking of an(implicit) promise.
Now in technology, the situation is arguably like this. Advertising a product as a coffee machine, adding to it an instruction manual showing you how its manipulation will result in a cup or a carafe of coffee, amounts to promising you that the product you bought will make coffee when plugged in (and not in a perverse way, i.e., by blowing up in your face while making coffee at the same time). An interesting case is one where conflicts of interests arise: where a promise is made that it is not, invariably, rational for the promise-maker to keep. It may be that corporate promises fall, ipso facto, in this category.Footnote 5 The fact that companies have introduced such things as guarantees may show that they realize that the reliability of the promises they make concerning their products is questionable. However, the contrast with the promises of individuals is easily exaggerated. Individuals are not less regularly tempted to reconsider their promises in the light of their own interests, and a promise without a reputation to support it does not easily lead to trust. The reason that a betrayal of trust is experienced as a significant event may well be that it represents a double failure: a failure on the part of the trustee to live up to a promissory expectation and a failure on the part of the truster to anticipate the contrary outcome. As the very word indignation indicates, the indignation felt when trust is betrayed is a combination of anger and shame.
Technical artifacts are in fact embedded in a network of promises: they are designed, manufactured, and sold as specimens of functional kinds; as things that are for being used in certain ways and thereby realizing certain outcomes, and that may therefore be expected to perform accordingly. These promises themselves, however, cannot ground trust concerning the artifacts’ operation. Many promises are made about second-hand cars, and they do not resolve the question whether the seller can be trusted; they make it more poignant. The promises make it the case that the disappointment of our trust concerning an artifact’s performance is also a betrayal of trust. In conditions where the producers seem not to be hindered by a bad reputation, and do not care about their reputation, as was typically the case under the former communist regimes, people are much less inclined to trust promises about artifact’s performance. If people are disappointed under such circumstances, it is disappointment about the failure of a plan, the sort of disappointment felt when a lottery ticket turns out not to have won a prize.
These considerations concern trusting the designers and manufacturers of technical artifacts, not the artifacts themselves. We could, however, introduce trust in artifacts in a derived sense. To trust an artifact to perform in a certain way would mean, then, to trust the designer or manufacturer of the artifact to have seen to it that the artifact performs, as we might put it, in accordance with the description of its function and only in accordance with the description of its function and operation, as communicated by the designer/manufacturer. If we trusted an artifact in this derived sense, we would consider it reliable.Footnote 6
From the perspective of this paper the claim that technical artifacts are trustworthy in this derived sense is not very interesting since it traces the trustworthiness of technical artifacts back to the interpersonal trustworthiness of designers, makers etc. What we are interested in here is whether the notion of trustworthiness may be applied to technical artifacts themselves, that is, whether a non-interpersonal notion of trustworthiness of technical artifacts may make sense. We have already concluded above that it is not at all clear how motivation-attributing accounts of trustworthiness may be adapted such that they may be applied to simple technical artifacts. In the following section we explore to what extent it is possible to apply motivation-attributing accounts to more complex technological objects of human reliance, in particular to technical objects that show agent-like behavior. We discuss the question whether they can show sufficient agent-like behavior to count as being in some sense trustworthy, and as suitable objects for trust. In such cases, however, we are dealing with ‘thin’ counterparts to the notions of trustworthiness and trust found within motivation-attributing accounts.