Open Questions in Creating Safe Open-ended AI: Tensions Between Control and Creativity

Artificial life originated and has long studied the topic of open-ended evolution, which seeks the principles underlying artificial systems that innovate continually, inspired by biological evolution. Recently, interest has grown within the broader field of AI in a generalization of open-ended evolution, here called open-ended search, wherein such questions of open-endedness are explored for advancing AI, whatever the nature of the underlying search algorithm (e.g. evolutionary or gradient-based). For example, open-ended search might design new architectures for neural networks, new reinforcement learning algorithms, or most ambitiously, aim at designing artificial general intelligence. This paper proposes that open-ended evolution and artificial life have much to contribute towards the understanding of open-ended AI, focusing here in particular on the safety of open-ended search. The idea is that AI systems are increasingly applied in the real world, often producing unintended harms in the process, which motivates the growing field of AI safety. This paper argues that open-ended AI has its own safety challenges, in particular, whether the creativity of open-ended systems can be productively and predictably controlled. This paper explains how unique safety problems manifest in open-ended search, and suggests concrete contributions and research questions to explore them. The hope is to inspire progress towards creative, useful, and safe open-ended search algorithms.

[1]  J. Hailman Wonderful Life: The Burgess Shale and the Nature of History, Stephen Jay Gould. W. W. Norton, New York (1989), 347, Price $19.95 (U.S.A.), $27.95 (Canada) , 1991 .

[2]  Thomas S. Ray,et al.  An Approach to the Synthesis of Life , 1991 .

[3]  M. Winston THE BIOLOGY AND MANAGEMENT OF AFRICANIZED HONEY BEES , 1992 .

[4]  Peter Danielson Artificial Morality: Virtuous Robots for Virtual Games , 1992 .

[5]  N. Packard,et al.  A Comparison of Evolutionary Activity in Artificial Evolving Systems and in the Biosphere , 1998 .

[6]  Tim Taylor,et al.  Replaying to tape: an investigation into the role of contingency in evolution , 1998 .

[7]  Russell K. Standish,et al.  Open-Ended Artificial Evolution , 2002, Int. J. Comput. Intell. Appl..

[8]  Laura H. Kahn Biodefense research: can secrecy and safety coexist? , 2004, Biosecurity and bioterrorism : biodefense strategy, practice, and science.

[9]  Christoph Adami,et al.  Influence of Chance, History, and Adaptation on Digital Evolution , 2004, Artificial Life.

[10]  Charles Ofria,et al.  Avida , 2004, Artificial Life.

[11]  Geoffrey Sayre-Mccord Moral realism , 2005 .

[12]  Ricardo Vilalta,et al.  A Perspective View and Survey of Meta-Learning , 2002, Artificial Intelligence Review.

[13]  C. Allen,et al.  Artificial Morality: Top-down, Bottom-up, and Hybrid Approaches , 2005, Ethics and Information Technology.

[14]  Huseyin Cavusoglu,et al.  Efficiency of Vulnerability Disclosure Mechanisms to Disseminate Vulnerability Knowledge , 2007, IEEE Transactions on Software Engineering.

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

[16]  Dario Floreano,et al.  Evolutionary Advantages of Neuromodulated Plasticity in Dynamic, Reward-based Scenarios , 2008, ALIFE.

[17]  Kenneth O. Stanley,et al.  Exploiting Open-Endedness to Solve Problems Through the Search for Novelty , 2008, ALIFE.

[18]  Melanie Mitchell,et al.  Complexity - A Guided Tour , 2009 .

[19]  N. Jacobstein Foresight Guidelines for Responsible Nanotechnology Development , 2010 .

[20]  R. Shine The Ecological Impact of Invasive Cane Toads (Bufo Marinus) in Australia , 2010, The Quarterly Review of Biology.

[21]  Kenneth O. Stanley,et al.  Novelty Search and the Problem with Objectives , 2011 .

[22]  C. Allen,et al.  Stanford Encyclopedia of Philosophy , 2011 .

[23]  Nick Bostrom,et al.  The Superintelligent Will: Motivation and Instrumental Rationality in Advanced Artificial Agents , 2012, Minds and Machines.

[24]  S. Armstrong General Purpose Intelligence: Arguing the Orthogonality Thesis , 2013 .

[25]  R. V. Schomberg A Vision of Responsible Research and Innovation , 2013 .

[26]  Robert Caverly,et al.  Responsible Innovation: Managing the Responsible Emergence of Science and Innovation in Society , 2013 .

[27]  Kenneth O. Stanley,et al.  A novel human-computer collaboration: combining novelty search with interactive evolution , 2014, GECCO.

[28]  Jason Yosinski,et al.  Innovation Engines: Automated Creativity and Improved Stochastic Optimization via Deep Learning , 2015, GECCO.

[29]  Kenneth O. Stanley,et al.  Open-Ended Evolution: Perspectives from the OEE Workshop in York , 2016, Artificial Life.

[30]  L. Trut Early Canid Domestication: The Farm-Fox Experiment Foxes hredfor tamability in a AO-year experiment exhibit remarkable transformations that suggest an interplay between behavioral genetics and development , 2016 .

[31]  H. P. de Vladar,et al.  Why Greatness Cannot Be Planned: The Myth of the Objective , 2016, Leonardo.

[32]  John Schulman,et al.  Concrete Problems in AI Safety , 2016, ArXiv.

[33]  Anca D. Dragan,et al.  Inverse Reward Design , 2017, NIPS.

[34]  Shane Legg,et al.  Deep Reinforcement Learning from Human Preferences , 2017, NIPS.

[35]  Kenneth O. Stanley,et al.  Minimal criterion coevolution: a new approach to open-ended search , 2017, GECCO.

[36]  Sergey Levine,et al.  Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks , 2017, ICML.

[37]  Arvind Satyanarayan,et al.  The Building Blocks of Interpretability , 2018 .

[38]  R. Lenski,et al.  Contingency and determinism in evolution: Replaying life’s tape , 2018, Science.

[39]  Owain Evans,et al.  Trial without Error: Towards Safe Reinforcement Learning via Human Intervention , 2017, AAMAS.

[40]  Marcus Hutter,et al.  AGI Safety Literature Review , 2018, IJCAI.

[41]  Risto Miikkulainen,et al.  The Surprising Creativity of Digital Evolution: A Collection of Anecdotes from the Evolutionary Computation and Artificial Life Research Communities , 2018, Artificial Life.

[42]  Shane Legg,et al.  Scalable agent alignment via reward modeling: a research direction , 2018, ArXiv.

[43]  Pieter Abbeel,et al.  Evolved Policy Gradients , 2018, NeurIPS.

[44]  Marcin Andrychowicz,et al.  Solving Rubik's Cube with a Robot Hand , 2019, ArXiv.

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

[46]  Jeff Clune,et al.  AI-GAs: AI-generating algorithms, an alternate paradigm for producing general artificial intelligence , 2019, ArXiv.

[47]  Jason Yosinski,et al.  Understanding Neural Networks via Feature Visualization: A survey , 2019, Explainable AI.

[48]  Kenneth O. Stanley,et al.  An Overview of Open-Ended Evolution: Editorial Introduction to the Open-Ended Evolution II Special Issue , 2019, Artificial Life.

[49]  Kenneth O. Stanley,et al.  POET: open-ended coevolution of environments and their optimized solutions , 2019, GECCO.

[50]  Nathaniel Virgo,et al.  On the Potential for Open-Endedness in Neural Networks , 2018, Artificial Life.

[51]  Jessica Taylor,et al.  Alignment for Advanced Machine Learning Systems , 2020, Ethics of Artificial Intelligence.