Skip to main content
SpringerLink
Log in

On the promotion of safe and socially beneficial artificial intelligence

  • Open Forum
  • Published:
AI & SOCIETY Aims and scope Submit manuscript

Abstract

This paper discusses means for promoting artificial intelligence (AI) that is designed to be safe and beneficial for society (or simply “beneficial AI”). The promotion of beneficial AI is a social challenge because it seeks to motivate AI developers to choose beneficial AI designs. Currently, the AI field is focused mainly on building AIs that are more capable, with little regard to social impacts. Two types of measures are available for encouraging the AI field to shift more toward building beneficial AI. Extrinsic measures impose constraints or incentives on AI researchers to induce them to pursue beneficial AI even if they do not want to. Intrinsic measures encourage AI researchers to want to pursue beneficial AI. Prior research focuses on extrinsic measures, but intrinsic measures are at least as important. Indeed, intrinsic factors can determine the success of extrinsic measures. Efforts to promote beneficial AI must consider intrinsic factors by studying the social psychology of AI research communities.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Similar content being viewed by others

Notes

  1. For perspectives on near-term AI impacts, see, e.g., Lin et al. (2011). For strong AI, see e.g., Eden et al. (2013).

  2. On the extrinsic/intrinsic distinction, see e.g., Markowitz and Sharif (2012:246) and references therein.

  3. See Sotala and Yampolskiy (2014, Sect. 3) for a review in the context of strong AI. Russell et al. (2015) also discuss a range of predominantly extrinsic measures. A notable exception to the focus on extrinsic measures is Russell’s emphasis on shifting “how practitioners define what they do” (Bohannon 2015:252).

  4. Incentives can nonetheless provoke significant backlash. For example, in the United States, environmentalists have long pursued incentive-based policies such as taxes on pollution in order to appeal to industry interests that do not want constraints, yet industry has been largely successful at avoiding these incentive-based policies.

  5. Incentives for completed technologies are less relevant for AIs that could be catastrophic because there may be no penalty that could adequately compensate for damages and, in the extreme case, no one alive to process the penalty.

  6. Conversely, when beneficial AI groups are identified, rewards are to be applied, though this is less of a challenge because AI groups are likely to seek rewards, not dodge them.

  7. Social sanctions are an extrinsic measure, specifically an incentive using a social penalty, though they can also cultivate certain social norms.

References

  • Amodei D, Olah C, Steinhardt J, Christiano P, Schulman J, Mané D (2016) Concrete problems in AI safety. arXiv:1606.06565

  • Armstrong S, Bostrom N, Shulman C (2016) Racing to the precipice: a model of artificial intelligence development. AI & Soc 31(2):201–206

    Article  Google Scholar 

  • Baum SD, Goertzel B, Goertzel TG (2011) How long until human-level AI? Results from an expert assessment. Technol Forecast Soc Change 78(1):185–195

    Article  Google Scholar 

  • Bohannon J (2015) Fears of an AI pioneer. Science 349(6245):252

    Article  MathSciNet  MATH  Google Scholar 

  • Borrie J (2014) Humanitarian reframing of nuclear weapons and the logic of a ban. Int Aff 90(3):625–646

    Article  Google Scholar 

  • Cooter RD (2000) Three effects of social norms on law: expression, deterrence, and internalization. Oregon Law Rev 79:1–22

    Google Scholar 

  • Deci EL (1971) Effects of externally mediated rewards on intrinsic motivation. J Pers Soc Psychol 18:105–115

    Article  Google Scholar 

  • Dickerson CA, Thibodeau R, Aronson E, Miller D (1992) Using cognitive dissonance to encourage water conservation. J Appl Soc Psychol 22:841–854

    Article  Google Scholar 

  • Eden AH, Moor JH, Soraker JH, Steinhart E (2013) Singularity hypotheses: a scientific and philosophical assessment. Springer, Berlin

    MATH  Google Scholar 

  • Griffith E (2016) Who will build the next great car company? Fortune Magazine, 23 June. http://fortune.com/self-driving-cars-silicon-valley-detroit. Accessed 26 Sept 2016

  • Gurney JK (2013) Sue my car not me: products liability and accidents involving autonomous vehicles. Univ Ill J Law Technol Policy 2013(2):247–277

    Google Scholar 

  • Joy B (2000) Why the future doesn’t need us. Wired 8(4):238–263

    Google Scholar 

  • Krantz DH, Peterson N, Arora P, Milch K, Orlove B (2008) Individual values and social goals in environmental decision making. In: Smith JC, Connolly T, Son YJ (eds) Kugler T. Decision modeling and behavior in complex and uncertain environments New York, Springer, pp 165–198

    Google Scholar 

  • Kunda Z (1990) The case for motivated reasoning. Psychol Bull 108(3):480–498

    Article  Google Scholar 

  • Lapinski MK, Rimal RN (2005) An explication of social norms. Commun Theory 15(2):127–147

    Article  Google Scholar 

  • Lewandowsky S, Cook J, Oberauer K, Brophy S, Lloyd EA, Marriott M (2015) Recurrent fury: conspiratorial discourse in the blogosphere triggered by research on the role of conspiracist ideation in climate denial. J Soc Political Psychol 3(1):142–178

    Article  Google Scholar 

  • Lin P, Abney K, Bekey GA (eds) (2011) Robot ethics: the ethical and social implications of robotics. MIT Press, Cambridge

    Google Scholar 

  • Markowitz EM, Shariff AF (2012) Climate change and moral judgement. Nat Clim Change 2(4):243–247

    Article  Google Scholar 

  • McGinnis JO (2010) Accelerating AI. Northwest Univ Law Rev 104:366–381

    Google Scholar 

  • Moses LB (2007) Recurring dilemmas: the law’s race to keep up with technological change. Univ Ill J Law Technol Policy 2007(2):239–285

    MathSciNet  Google Scholar 

  • Posner EA (2000) Law and social norms: the case of tax compliance. Va Law Rev 86:1781–1819

    Article  Google Scholar 

  • Posner RA (2004) Catastrophe: risk and response. Oxford University Press, Oxford

  • Russell S, Dewey D, Tegmark M (2015) Research priorities for robust and beneficial artificial intelligence. AI Mag 36(4):105–114

    Article  Google Scholar 

  • Schienke EW, Tuana N, Brown DA, Davis KJ, Keller K, Shortle JS, Stickler M, Baum SD (2009) The role of the NSF Broader Impacts Criterion in enhancing research ethics pedagogy. Soc Epistemol 23(3–4):317–336

    Article  Google Scholar 

  • Schienke EW, Baum SD, Tuana N, Davis KJ, Keller K (2011) Intrinsic ethics regarding integrated assessment models for climate management. Sci Eng Ethics 17(3):503–523

    Article  Google Scholar 

  • Schultz PW, Nolan JM, Cialdini RB, Goldstein NJ, Griskevicius V (2007) The constructive, destructive, and reconstructive power of social norms. Psychol Sci 18(5):429–434

    Article  Google Scholar 

  • Scruggs L, Benegal S (2012) Declining public concern about climate change: can we blame the great recession? Glob Environ Change 22(2):505–515

    Article  Google Scholar 

  • Shome D, Marx S (2009) The psychology of climate change communication: a guide for scientists, journalists, educators, political aides, and the interested public. Columbia University Center for Research on Environmental Decisions, New York

    Google Scholar 

  • Shulman C (2009) Arms control and intelligence explosions. In: 7th European Conference on Computing and Philosophy (ECAP), Bellaterra, Spain, July 2–4

  • Sotala K, Yampolskiy RV (2014) Responses to catastrophic AGI risk: a survey. Physica Scripta 90(1):018001. doi:10.1088/0031-8949/90/1/018001

  • Stone J, Fernandez NC (2008) To practice what we preach: the use of hypocrisy and cognitive dissonance to motivate behavior change. Soc Pers Psychol Compass 2(2):1024–1051

    Article  Google Scholar 

  • Sunstein CR (1996) On the expressive function of law. Univ Pa Law Rev 144(5):2021–2053

    Article  Google Scholar 

  • Vohs KD, Mead NL, Goode MR (2006) The psychological consequences of money. Science 314:1154–1156

    Article  Google Scholar 

  • Wilson G (2013) Minimizing global catastrophic and existential risks from emerging technologies through international law. Va Environ Law J 31:307–364

    Google Scholar 

  • Yampolskiy R, Fox J (2013) Safety engineering for artificial general intelligence. Topoi 32(2):217–226

    Google Scholar 

Download references

Acknowledgments

This paper has benefited from conversations with many people including Dario Amodei, Miles Brundage, Sean Legassick, Richard Mallah, and Jaan Tallinn, and from time spent at the Columbia University Center for Research on Environmental Decisions. Tony Barrett, Steven Umbrello, and Peter Howe provided helpful comments on an earlier draft. Any errors or shortcomings in the paper are the author’s alone. Work on this paper was funded in part by a grant from the Future of Life Institute. The views in this paper are the author’s and are not necessarily the views of the Future of Life Institute or the Global Catastrophic Risk Institute.

Author information

Authors and Affiliations

  1. Global Catastrophic Risk Institute, PO Box 40364, Washington, DC, 20016, USA

    Seth D. Baum

Authors
  1. Seth D. Baum
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Seth D. Baum.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Baum, S.D. On the promotion of safe and socially beneficial artificial intelligence. AI & Soc 32, 543–551 (2017). https://doi.org/10.1007/s00146-016-0677-0

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00146-016-0677-0

Keywords

Search

Navigation