In-Group or Out-Group? A Role for Living Machines in Human Society

Transcription

1 In-Group or Out-Group? A Role for Living Machines in Human Society Joanna J. Bryson Artificial Models of Natural Intelligence University of Bath, United Kingdom Mannheimer Zentrum für Europäische Sozialforschung

2 In Groups, Out Groups & Public Goods Figure: Hermann, Thöni & Gächter (2008) The value of a specific public good varies by context. It is possible to over-invest in one group s assets. Joanna J. Bryson, James Mitchell, Simon T. Powers and Karolina Sylwester, Understanding and Addressing Cultural Variation in Costly Antisocial Punishment, in Applied Evolutionary Anthropology, M. A. Gibson and D. W. Lawson (eds.), Springer, in press. Karolina Sylwester, Benedikt Herrmann and Joanna J. Bryson, Homo homini lupus? Anti-social punishment revisited, Journal of Neuroscience, Psychology, and Economics, in press.

3 Robot Ethics Robots are servants we own. Slaves We build them and determine their goals, either directly or by specifying how they acquire goals themselves. Joanna J. Bryson, Robots Should Be Slaves in Close Engagements with Artificial Companions: Key social, psychological, ethical and design issues, Yorick Wilks (ed.), pp , John Benjamins, Amsterdam, 2010.

4 icub, January 2010 Genoa (ROBOTCUB Final Review Meeting) photos: me

5 Because we build robots and determine their goals, our relationship to them is fundamentally different from our relationship to humans or other evolved systems. photos: Georgio Metta (top) & Emmanuel Tanguy

6 We Get to Decide... Whether robot minds are unique or have autosave and offsite backup. Whether robots suffer (permanently change their behaviour) when they are neglected, insulted or constrained. Whether robots compete with us (or our customers) for time, money & materials.

7 Reasons some might object to robot slavery Sci Fi Diesel Sweeties Science fiction uses alien perspectives to examine ourselves by analogy. Works via establishing empathy. Fiction. HHGTTG TV series

8 Reasons others might object to robot slavery Liability Since 2008 there have been more robots in Iraq than non-usa coalition forces. Powerful incentives for corporate & political liability to be transferred. Joanna J. Bryson Why Robot Nannies ProbablyWon t Do Much Psychological Damage, commentary on Noel Sharkey and Amanda Sharkey, The Crying Shame of Robot Nannies: An Ethical Appraisal, Interaction Studies, 11(2): , June 2010.

9 People want to make AI they owe obligations to, can fall in love with, etc. equals over which we have complete dominion. Moral hazard constructing artefacts that can demand resources or be ascribed blame. Joanna J. Bryson and Philip P. Kime, Just an Artifact: Why Machines are Perceived as Moral Agents, The Twenty-Second International Joint Conference on Artificial Intelligence (IJCAI), Barcelona, Spain, pp , Morgan Kaufmann, Joanna J. Bryson Patiency Is Not a Virtue: Suggestions for Co-Constructing an Ethical Framework Including Intelligent Artefacts, The Machine Question: AI, Ethics and Moral Responsibility, D. Gunkel, S. Torrance and J. J. Bryson (eds.), AISB, Birmingham, 2012.

10 Slaves Corollary I We (as policy makers, academics and manufacturers) are obliged to make sure that we (as humans) have no obligation to robots. For example, robot minds / memories should be backed up continuously in real time. Joanna J. Bryson, Robots Should Be Slaves in Close Engagements with Artificial Companions: Key social, psychological, ethical and design issues, Yorick Wilks (ed.), pp , John Benjamins, Amsterdam, Joanna Bryson, A Proposal for the Humanoid Agent-builders League (HAL), AISB 00 Symposium on Artificial Intelligence, Ethics and (Quasi-)Human Rights, John Barnden (ed.), pp. 1 6; Birmingham UK, 2000.

11 Slaves Corollary II We (as policy makers, academics and manufacturers) should not deceive people into misallocating resources to robots, even if it makes us money or wins us votes. Examples of resources: time, money, attention, personal care.

12 UK EPSRC s Principles of Robotics (2011) 1. Robots are multi-use tools. Robots should not be designed solely or primarily to kill or harm humans, except in the interests of national security. 2. Humans, not robots, are responsible agents. Robots should be designed & operated as far as is practicable to comply with existing laws & fundamental rights & freedoms, including privacy. 3. Robots are products. They should be designed using processes which assure their safety and security. (of 5...)

13 UK EPSRC s Principles of Robotics (2011) 4. Robots are manufactured artefacts. They should not be designed in a deceptive way to exploit vulnerable users; instead their machine nature should be transparent. 5. The person with legal responsibility for a robot should be attributed. [like automobile titles] Joanna J. Bryson, Kerstin Dautenhahn and Geoff Pegman, Man and the machine, letter published online, The Economist, 16 June Joanna J. Bryson The Making of the EPSRC Principles of Robotics, The AISB Quarterly, (133) Spring 2012.

14 Main Message AI ethics relates two types of human artefact: ethical systems & robots. Normative not descriptive ethics: there is no pre-determined social place for AI. Question: is there any utility in displacing the responsibility we as authors have onto AI? Not a question: is it possible. is does not imply ought Joanna J. Bryson Patiency Is Not a Virtue: Suggestions for Co-Constructing an Ethical Framework Including Intelligent Artefacts, in The Machine Question: AI, Ethics and Moral Responsibility, D. Gunkel, S. Torrance and J. J. Bryson (eds.), AISB, Birmingham, UK, 2012.

15 (1998) AI is Just An(other) Artifact We misallocate fear & hope because we identify with it. This leads us to ignore real hazards. Phil Kime (2012)

16 Humans Are Just Chimps With History A brain 3x the size of a chimp s is not a big deal. Human uniqueness derives from our rate of communication (Bryson 2009, 2010). We only became really unique in the last 10,000 years, with writing, urbanisation, doctrinal religion & agriculture.

17 AI Has Our Advantage big dog Watson driverless cars Big data + Bayesian-based machine learning has successfully mined human intelligence in a broad range of domains.

18 Moral Patiency Is not a Virtue AI ethics relates two types of human artefact: ethical systems & robots. Normative not descriptive ethics: there is no pre-determined social place for AI. Question: is there any utility in displacing the responsibility we as authors have onto AI? Not a question: is it possible.

19 Moral Actions a behavioural context affords more than one possible action for the individual, at least one available action is considered by a society to be more socially beneficial than the other options, and the individual is able to recognise which action is socially sanctioned and act on this information. Easy to build in robots!! (includes monkeys & pets!)

20 Displacement of Responsibility What are the pros & cons of considering robots moral subjects? at least one available action is considered by a society to be more socially beneficial than the other options, and

21 For Human Society (us): Pros: feel godlike, culture might persist beyond planetary limits, might produce more useful tools. Cons: political & commercial moral hazard, misattribution of blame / resources. For AI (them robots): No Pros: (except maybe for the unbuilt). Cons: compete w/ humans for resources, stress of social dominance, fear of death etc.

22 Conclusion Organising our society such that AI is a moral agent or patient benefits no one. Bad for humans; bad for machines.

23 Thanks! Phil Kime, Will Lowe, David Gunkel, my PhD students & other collaborators, and you for your attention.