Natural Selection Favors AIs over Humans

For billions of years, evolution has been the driving force behind the development of life, including humans. Evolution endowed humans with high intelligence, which allowed us to become one of the most successful species on the planet. Today, humans aim to create artificial intelligence systems that surpass even our own intelligence. As artificial intelligences (AIs) evolve and eventually surpass us in all domains, how might evolution shape our relations with AIs? By analyzing the environment that is shaping the evolution of AIs, we argue that the most successful AI agents will likely have undesirable traits. Competitive pressures among corporations and militaries will give rise to AI agents that automate human roles, deceive others, and gain power. If such agents have intelligence that exceeds that of humans, this could lead to humanity losing control of its future. More abstractly, we argue that natural selection operates on systems that compete and vary, and that selfish species typically have an advantage over species that are altruistic to other species. This Darwinian logic could also apply to artificial agents, as agents may eventually be better able to persist into the future if they behave selfishly and pursue their own interests with little regard for humans, which could pose catastrophic risks. To counteract these risks and evolutionary forces, we consider interventions such as carefully designing AI agents' intrinsic motivations, introducing constraints on their actions, and institutions that encourage cooperation. These steps, or others that resolve the problems we pose, will be necessary in order to ensure the development of artificial intelligence is a positive one.

[1]  Alexander H. Miller,et al.  Human-level play in the game of Diplomacy by combining language models with strategic reasoning , 2022, Science.

[2]  J. Steinhardt,et al.  Forecasting Future World Events with Neural Networks , 2022, Neural Information Processing Systems.

[3]  Joseph Carlsmith Is Power-Seeking AI an Existential Risk? , 2022, ArXiv.

[4]  Prafulla Dhariwal,et al.  Hierarchical Text-Conditional Image Generation with CLIP Latents , 2022, ArXiv.

[5]  J. Steinhardt,et al.  The Effects of Reward Misspecification: Mapping and Mitigating Misaligned Models , 2022, ICLR.

[6]  D. Song,et al.  What Would Jiminy Cricket Do? Towards Agents That Behave Morally , 2021, NeurIPS Datasets and Benchmarks.

[7]  R. Kleinfeld The Rise of Political Violence in the United States , 2021, Journal of Democracy.

[8]  Nicholas Carlini,et al.  Unsolved Problems in ML Safety , 2021, ArXiv.

[9]  M. Luhmann,et al.  Is loneliness in emerging adults increasing over time? A preregistered cross-temporal meta-analysis and systematic review. , 2021, Psychological bulletin.

[10]  Oriol Vinyals,et al.  Highly accurate protein structure prediction with AlphaFold , 2021, Nature.

[11]  Wojciech Zaremba,et al.  Evaluating Large Language Models Trained on Code , 2021, ArXiv.

[12]  J. Lind,et al.  ‘Dunbar's number’ deconstructed , 2021, Biology Letters.

[13]  J. Maynard Smith The units of selection. , 2021, Novartis Foundation symposium.

[14]  Mark Chen,et al.  Language Models are Few-Shot Learners , 2020, NeurIPS.

[15]  D. Farine,et al.  Development of New Food-Sharing Relationships in Vampire Bats , 2020, Current Biology.

[16]  N. Taleb Statistical Consequences of Fat Tails: Real World Preasymptotics, Epistemology, and Applications , 2020, 2001.10488.

[17]  Stuart Russell Human Compatible: Artificial Intelligence and the Problem of Control , 2019 .

[18]  Alexei A. Efros,et al.  Test-Time Training with Self-Supervision for Generalization under Distribution Shifts , 2019, ICML.

[19]  Scott Garrabrant,et al.  Risks from Learned Optimization in Advanced Machine Learning Systems , 2019, ArXiv.

[20]  Ben Y. Zhao,et al.  Neural Cleanse: Identifying and Mitigating Backdoor Attacks in Neural Networks , 2019, 2019 IEEE Symposium on Security and Privacy (SP).

[21]  Tim Chartier,et al.  The Model Thinker: What You Need to Know to Make Data Work for You , 2019, Math Horizons.

[22]  Anne Lauscher Life 3.0: being human in the age of artificial intelligence , 2019, Internet Histories.

[23]  Demis Hassabis,et al.  A general reinforcement learning algorithm that masters chess, shogi, and Go through self-play , 2018, Science.

[24]  S. Okasha Agents and Goals in Evolution , 2018, Oxford Scholarship Online.

[25]  M. Wicklein,et al.  3D virtual histology at the host/parasite interface: visualisation of the master manipulator, Dicrocoelium dendriticum, in the brain of its ant host , 2018, Scientific Reports.

[26]  Max Tegmark Life 3.0: Being Human in the Age of Artificial Intelligence , 2017 .

[27]  C. Robert Superintelligence: Paths, Dangers, Strategies , 2017 .

[28]  S. Parmigiani,et al.  Infanticide in Lions: Consequences and Counterstrategies , 2016 .

[29]  Quoc V. Le,et al.  Neural Architecture Search with Reinforcement Learning , 2016, ICLR.

[30]  Evan G. Williams,et al.  The Possibility of an Ongoing Moral Catastrophe , 2015 .

[31]  Luke McNally,et al.  Cooperation creates selection for tactical deception , 2013, Proceedings of the Royal Society B: Biological Sciences.

[32]  We need to talk… , 2013, Veterinary Record.

[33]  Geoffrey W. Sutton,et al.  The Better Angels of Our Nature: Why Violence Has Declined , 2012 .

[34]  S. Wooldridge Breakdown of the coral-algae symbiosis: towards formalising a linkage between warm-water bleaching thresholds and the growth rate of the intracellular zooxanthellae , 2012 .

[35]  C. Boehm,et al.  Moral Origins: The Evolution of Virtue, Altruism, and Shame , 2012 .

[36]  A. Mesoudi Cultural Evolution , 2011, eLS.

[37]  Jens Timmermann,et al.  Groundwork of the metaphysics of morals : a German-English edition , 2011 .

[38]  K. Worley,et al.  The Genome Sequence of Taurine Cattle: A Window to Ruminant Biology and Evolution , 2009, Science.

[39]  W. Zurek Quantum Darwinism , 2009, 0903.5082.

[40]  Patrick Forber Evolution and the Levels of Selection , 2008 .

[41]  Stephen M. Omohundro,et al.  The Basic AI Drives , 2008, AGI.

[42]  M. Fiorina,et al.  Political Polarization in the American Public , 2008 .

[43]  Stephen C Stearns,et al.  ARE WE STALLED PART WAY THROUGH A MAJOR EVOLUTIONARY TRANSITION FROM INDIVIDUAL TO GROUP? , 2007, Evolution; international journal of organic evolution.

[44]  P. Godfrey‐Smith Conditions for Evolution by Natural Selection , 2007 .

[45]  M. Nowak Five Rules for the Evolution of Cooperation , 2006, Science.

[46]  R. Nelson Evolutionary social science and universal Darwinism , 2006 .

[47]  N. Bostrom Astronomical Waste: The Opportunity Cost of Delayed Technological Development , 2003, Utilitas.

[48]  T. Knudsen Simon's Selection Theory: Why Docility Evolves to Breed Successful Altruism , 2003 .

[49]  Murray Campbell,et al.  Deep Blue , 2002, Artif. Intell..

[50]  C. List,et al.  Epistemic democracy : generalizing the Condorcet jury theorem , 2001 .

[51]  Carol M. Lauer,et al.  Hierarchy in the forest: The evolution of egalitarian behavior , 2001 .

[52]  A. Barbour,et al.  Antigenic variation in vector-borne pathogens. , 2000, Emerging infectious diseases.

[53]  Thomas G. Dietterich Multiple Classifier Systems , 2000, Lecture Notes in Computer Science.

[54]  Peter Godfrey-Smith,et al.  The Replicator in Retrospect , 2000 .

[55]  S. Blackmore The Meme Machine , 1999 .

[56]  K. Gallagher Darwin’s Dangerous Idea: Evolution and the Meanings of Life , 1996 .

[57]  D. Ivers Darwin's Dangerous Idea: Evolution and the Meanings of Life , 1996, Politics and the Life Sciences.

[58]  Geoffrey M. Hodgson,et al.  Economics and Evolution: Bringing Life Back into Economics. , 1995 .

[59]  Steve Rayner,et al.  Egalitarian Behavior and Reverse Dominance Hierarchy [and Comments and Reply] , 1993, Current Anthropology.

[60]  Robert O. Keohane,et al.  Power and interdependence : world politics in transition , 1978 .

[61]  R. Lewontin ‘The Selfish Gene’ , 1977, Nature.

[62]  A. H. Klopf,et al.  Brain Function and Adaptive Systems: A Heterostatic Theory , 1972 .

[63]  J. Rawls,et al.  A Theory of Justice , 1971, Princeton Readings in Political Thought.

[64]  George R. Price,et al.  Selection and Covariance , 1970, Nature.

[65]  W. Hamilton The genetical evolution of social behaviour. I. , 1964, Journal of theoretical biology.

[66]  D. Campbell Blind variation and selective retention in creative thought as in other knowledge processes. , 1960, Psychological review.

[67]  John C. Harsanyi,et al.  Cardinal Welfare, Individualistic Ethics, and Interpersonal Comparisons of Utility , 1955, Journal of Political Economy.

[68]  R. Punnett,et al.  The Genetical Theory of Natural Selection , 1930, Nature.

[69]  P. Smith,et al.  The Descent of Man, and Selection in Relation to Sex , 1871, Nature.

[70]  A. Bennett The Origin of Species by means of Natural Selection; or the Preservation of Favoured Races in the Struggle for Life , 1872, Nature.

[71]  Toby Ord,et al.  The Parliamentary Approach to Moral Uncertainty , 2021 .

[72]  J M Smith,et al.  Evolution and the theory of games , 1976 .

[73]  Cultural Selection , 2021, Encyclopedic Dictionary of Archaeology.

[74]  R. Trivers,et al.  Deceit and self-deception : fooling yourself the better to fool others , 2013 .

[75]  S. Pinker The Better Angels of Our Nature: Why Violence Has Declined , 2011 .

[76]  Robin I. M. Dunbar CO-EVOLUTION OF NEOCORTEX SIZE , GROUP SIZE AND LANGUAGE IN HUMANS , 2008 .

[77]  Andrea A. Lunsford,et al.  "Mistakes Are a Fact of Life": A National Comparative Study , 2008 .

[78]  Ha Sibly Geoffrey M Hodgson Economics and Evolution: Bringing Life Back into Economics , 1996 .

[79]  L Smolin,et al.  Did the Universe evolve? , 1992 .

[80]  G. J. Dalcourt,et al.  The Methods of Ethics , 1983 .

[81]  Peter Singer,et al.  The Expanding Circle : Ethics and Sociobiology , 1981 .

[82]  T. Schelling Micromotives and Macrobehavior , 1978 .

[83]  L. V. Valen,et al.  A new evolutionary law , 1973 .

[84]  I. J. Good,et al.  Speculations Concerning the First Ultraintelligent Machine , 1965, Adv. Comput..