The First Law of Robotics (A Call to Arms)

By Daniel S. Weld, Oren Etzioni. 1994

Subfields: Safety, Robustness & Interpretability

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Safety, Robustness & Interpretability

Citations in the corpus

Scalable agent alignment via reward modeling: a research direction (2018, background): “Since then, the agent alignment problem has been echoed by philosophers (Bostrom, 2003, 2014; Yudkowsky, 2004) and treated informally by technical authors (Wiener, 1960; Etzioni & Weld, 1994; Omohundro, 2008).”
Research Priorities for Robust and Beneficial Artificial Intelligence (2015, background): “has therefore stoked interest in high-assurance systems where strong robustness guarantees can be made; Weld and Etzioni have argued that ‘‘society will reject autonomous agents unless we have some credible means of making them safe” (Weld and Etzioni 1994).”
AI Safety Gridworlds (2017, background): “Explicitly specifying all such safety constraints (e.g. Weld and Etzioni, 1994) is both labor-intensive and brittle, and unlikely to scale or generalize well.”
AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation Framework (2023, background): “mode AutoGen supports could be desirable in many scenarios, a high level of autonomy can also pose potential risks, especially in high-risk/stakes applications [4, 50].”
Agent Foundations for Aligning Machine Intelligence with Human Interests: A Technical Research Agenda (2017, background): “This is our call to arms: regardless of whether research efforts follow the path laid out in this document, significant effort must be focused on the study of superintelligence alignment as soon as possible.”
Assuring the Behavior of Adaptive Agents (2006, background): “This law, which has also been studied by Weld and Etzioni [ 34 ], states that an agent may not harm a human or allow a human to come to harm.”
PDDL-the planning domain definition language (1998, background): “For details of safety constraints, please refer to [11].”
AutoGen: Enabling Next-Gen LLM Applications via Multi-Agent Conversation (2023, unknown): “While the autonomous mode AutoGen supports could be desirable in many scenarios, a high level of autonomy can also pose potential risks, especially in high-risk applications (Amodei et al., 2016; Weld & Etzioni, 1994).”
Using temporal logics to express search control knowledge for planning (2000, background): “Such goals generalize the safety and maintenance goals mentioned in [60].”
Aligning Superintelligence with Human Interests: A Technical Research Agenda (2015, background): “Weld and Etzioni (1994) directed a “call to arms” to computer scientists, noting that “society will reject autonomous agents unless we have some credible means of making them safe.””
Should Robots be Obedient? (2017, background): “The suggestion that it might sometimes be better for an autonomous systems to be disobedient is not new [Weld and Etzioni, 1994; Scheutz and Crowell, 2007].”
Automated Planning and Acting (2016, background): “The UCPOP planner [472, 38, 599] extended SNLP to handle some extensions to the classical domain representation, including conditional effects and universally quantified effects [469, 470] Several other extensions have also been studied, such as incomplete information and sensing actions [474, 184, ”
Scruples: A Corpus of Community Ethical Judgments on 32, 000 Real-Life Anecdotes (2020, background): “In the intervening years, many researchers cautioned about super-intelligence and the need for AI to understand ethics (Vinge 1993; Weld and Etzioni 1994).”
Lyapunov Design for Safe Reinforcement Learning (2003, background): “Practitioners of artificial intelligence are paying increasing attention to the reliability and safety of approximate solutions to sequential decision problems (Singh et al., 1994; Weld and Etzioni, 1994; Schneider, 1997; Neuneier and Mihatsch, 1999; Gordon, 2000; Perkins and Barto, 2001a,b).”