What responsibilities should AI creators and firms have for the behavior and impact of their models, and why?

Most answers emphasize accountability, legal compliance, and risk mitigation for AI creators, though some highlight transparency, fairness, and the need for external oversight or self-regulation.

12 answers

Answers

  • Dean Ball

    Head of Strategic Futures · Openai

    Accountability for Misuse

    Dean Ball says AI developers should be accountable for unsafe development practices and that people and firms using AI should be held accountable when their misuse or lack of oversight causes harm, arguing this requires legal and technical infrastructure tying people to their agents' actions.

    Don't Worry About the Vase · Don't Worry About the Vase · Read
  • Ali Partovi

    CEO · Neo

    Hold them as if intentional

    Ali Partovi says the law should hold AI companies responsible when their AI does something illegal and treat it as if they did it intentionally, at the highest penalty, because treating such incidents as accidents keeps them happening.

    MTS · Listen on Apple Podcasts
  • Reid Hoffman

    Co-Founder, Executive Board Chair · Manas Ai

    Ensure accountability

    Reid Hoffman says figuring out accountability and accountability laws is important, and that the right place to start is to commission smart work on a dynamic framework for evolving accountability.

    Reid Hoffman · Watch on YouTube
  • Tyler Cowen

    Professor · George Mason University

    Make models publicly available

    Tyler Cowen says that if companies use powerful models internally for research or engineering, they should also make them publicly available, because he thinks good decisions can't be made without access to these models.

    Interesting Times with Ross Douthat · Watch on YouTube
  • Daron Acemoglu

    Institute Professor · Massachusetts Institute of Technology

    Ensure fairness and compensate data contributors

    Daron Acemoglu says firms should compensate the workers who generate their data, calling that fairness and also a good business decision.

    The answer starts at 41:46.

    Silicon Valley Girl: AI, Tech and Career Growth · Listen on Apple Podcasts
  • Tom Bilyeu

    CEO · Impact Theory

    Ensure Safety

    Tom Bilyeu says AI company leaders like Dario have a duty to employees and shareholders to make the company successful but also a responsibility to do things safely.

    Tom Bilyeu's Impact Theory · Tom Bilyeu's Impact Theory · Listen at 6:22
  • Vlad Tenev

    Co-Founder, Chairman, CEO · Robinhood

    Be Held Responsible

    Vlad Tenev does not give this answer in the passage; the quoted remarks about liability and being held responsible come from Treasury Secretary Scott Bessent, not Tenev.

    The answer starts at 4:25.

    Moonshots with Peter Diamandis · Listen on Apple Podcasts
  • Dario Amodei

    CEO and Co-Founder · Anthropic

    Add external evaluators and coordinate limits

    Dario Amodei proposes that AI firms embed external evaluators free to publish their findings uncensored and that democratic countries coordinate to set common limits, as steps to slow the pace of AI development.

    Quality Value · Quality Value · Read
  • Richard Socher

    CEO · Recursive

    Ensure legal compliance

    Richard Socher says AI must ultimately be aligned with the law wherever it is deployed, framing legal alignment as the baseline requirement for AI creators.

    Latent Space · Latent Space · Read
  • Greg Brockman

    President, Chairman, & Co-Founder · Openai

    Mitigate risks

    Greg Brockman says AI creators must take the impacts and risks seriously and really mitigate them, while also communicating the nuance so people don't throw out the baby with the bathwater.

    The answer starts at 37:39.

    Odd Lots · Listen on Apple Podcasts
  • Anders Hejlsberg

    Technical Fellow · Microsoft

    Take responsibility

    Anders Hejlsberg says someone has to take responsibility for what AI is doing, and if you give it the keys, it's on you, because users will sue you rather than the AI.

    Ryan Peterman · Watch on YouTube
  • Rhett Alden

    CTO, Elsevier Health Markets · Elsevier

    Self‑regulate responsibly

    Rhett Alden says his company, Elsevier, is establishing its own responsible AI principles and trust factors rather than waiting for regulators, because regulation can't keep up with AI's pace and self-regulation is currently the norm.

    The answer starts at 5:32.

    The AI in Business Podcast · Listen on Apple Podcasts

More Generative AI questions

View all →