Brief IA
›

AI: Steven Pinker Advocates for a Measured Approach to Catastrophic Scenarios

Regulation & Ethics·Tom Levy·

AI: Steven Pinker Advocates for a Measured Approach to Catastrophic Scenarios

AI: Steven Pinker Advocates for a Measured Approach to Catastrophic Scenarios
Key Takeaways
1Steven Pinker rejects apocalyptic scenarios and advocates for AI safety based on independent oversight, accountability, and human control.
2He admits to having underestimated the capabilities of large language models and criticizes the haste of certain product launches.
3Scott Alexander estimates the risk of extinction related to AI at 20% and calls for slowing down development through international agreements.
Why it matters : The debate contrasts a pragmatic approach to safety with a quantified assessment of existential risk, leading to divergent recommendations for regulating AI.

Harvard psychologist Steven Pinker dismisses scenarios of annihilation by AI and advocates for safety engineering based on independent oversight, accountability, and human control. Scott Alexander, to whom the letter is addressed, estimates the risk of extinction at 20% and calls for a slowdown in development.

Targeted Risks and Proposed Responses by Pinker and Alexander

Steven Pinker views bioterrorism, cyberattacks, and uncontrolled AI agents as real dangers. He believes that apocalyptic rhetoric does not allow for effective handling of these issues and recommends a sober approach to safety engineering, relying on independent oversight, accountability, and human control. Scott Alexander, who presented his position in June, estimates the probability that AI could annihilate humanity at 20% and wishes to strengthen security research as well as establish international agreements to slow down AI development.

Le brief IA que lisent les pros

Le brief IA que les pros lisent chaque soir

Les 7 actus IA du jour, décryptées en 5 min. Gratuit.

Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.

Choisissez votre rythme

Gratuit · Pas de spam · Désabonnement en 1 clic

Disagreement on Catastrophic AI Scenarios

Steven Pinker challenges apocalyptic scenarios such as that of the paperclip maximizer, which he sees as a confusion between intelligence and a desire for domination. He emphasizes that these narratives attribute to AI an obsession with a single goal and considers fears of an AI capable of exterminating humanity to be exaggerated.

A Refused Public Exchange and a Revision of Judgment

Steven Pinker publicly released his open letter on Quillette, addressed to Scott Alexander, an influential psychiatrist and tech blogger who had proposed a public debate on the existential risks associated with AI. Pinker declined this proposal, describing such debates as a "spectator sport." He indicates that he has had an optimistic view of AI but admits to having underestimated the performance of large language models and not foreseeing the speed with which certain companies commercialized their products, a speed he attributes to members of the same community that warns against these risks. Pinker practices as a psychologist at Harvard.

Brief IA — L'actualité IA en français

L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.