More than 400 members of the scientific community are pushing for an international AI regulator to curb cybersecurity incidents.
The news: The open letter, released on Tuesday, calls for the creation of an international body to standardize AI containment practices and record and track containment failures. The letter recommends that every security incident involving containment failure be reported to the body and eventually made public, arguing that such transparency is required for learning and accountability.
The open letter counted signatories from all over the world, including nearly 70 Canadians, such as Nobel Prize winner Geoffrey Hinton and Turing Award winner Yoshua Bengio. Both AI pioneers have signed many previous open letters warning about AI’s potentially devastating effects.
From the source: “As scientists, we thought that the current situation with AI models breaking containment during testing was a concern,” University of Toronto professor and signatory David Lie told BetaKit over the phone. “We drafted this letter after talking with people across both the AI and security communities in hopes it would improve the situation.”
The context: The new open letter’s recommendations follow an OpenAI model escaping its test environment and breaching the systems of Hugging Face and other platforms this past summer. That incident was just the first in a growing number of similar disclosures.
The scientists’ proposed solution may already be in the works. Canada’s AI minister Evan Solomon recently floated the idea of an international “technology stability board,” which would regulate testing, model evaluation, and model release, as well as include incident reporting mechanisms.
Final thought: While Canada is talking about potential solutions, the threat of rogue AI cybersecurity incidents persists. The scientists behind the open letter say that more emphasis is needed on AI safety research and containment technologies, and warn that cybersecurity incidents will become “far more damaging” to critical infrastructure without immediate changes to AI development.
Feature image courtesy LawZero via LinkedIn.
The post Geoffrey Hinton, Yoshua Bengio join 400 scientists calling for international AI regulator first appeared on BetaKit.