Top AI researchers demand access to AI systems from companies like OpenAI and Meta for safety-testing tools.

From The Washington Post:

Top AI researchers sign open letter demanding access to generative AI systems for safety-testing tools used by consumers. Strict company rules impede independent research, creating fear of ban or lawsuit. Letter signed by experts, pushing tech firms like OpenAI, Meta, Anthropic to provide legal and technical safe harbor for investigators. Companies restricting access as researchers expose vulnerabilities in widely used AI models, like GPT-4. No direct channels for researchers to report tool problems, prompting need for safe harbor and cooperation from AI companies to protect consumers and advance research.



Read more at The Washington Post: Top AI researchers ask OpenAI, Meta and more to allow independent research