Behind AI Industry’s Call for Independent Reviews, a Fight over What Gets Made Public: Reuters Exclusive
The companies are weighing redactions and approval rights as evaluators push for independent findings and broader access to model testing.
- Anthropic and OpenAI are negotiating strict publication rights for external AI safety evaluators, seeking significant editorial control over findings to prevent disclosure of trade secrets.
- AI labs fear embedded evaluators could inadvertently expose proprietary intellectual property to competitors, while experts argue the labs use access as leverage to control what gets reported.
- More than 200 researchers, including Geoffrey Hinton, signed a letter last month demanding independent control of findings after OpenAI redacted information from a Model Evaluation and Threat Research and Redwood Research investigation.
- Faculty CEO Marc Warner told Reuters, "Evaluators are only useful if people can trust what they report, so independence and transparency are critical." Without agreed standards, experts warn labs may dictate how systems are examined.
- President Donald Trump and leading executives signed the White House Accord on Super Intelligence last month, but the agreement, which Trump called "morally binding," did not specify how reviews should be conducted.
19 Articles
19 Articles
External artificial intelligence evaluators are specialized organizations or consultants that analyze AI models before or during their launch to detect security risks, inappropriate behaviors or vulnerabilities. However, their independence has become a point of friction between these entities and major technology companies. According to a research by Reuters, OpenAI and Anthropic want to exercise "significant control" over what external evaluato…
Developers try to hide the vulnerability of neuronets? Anthropic and OpenAI want to control the materials of independent experts who verify the safety of their IE models. Companies fear that the publication of such documents may lead to the disclosure of commercial secrets and technological developments. This is written by RBK-Ukraine with reference to Reuters. Disputes over the right to publish cause developers fear that the auditors involved m…
AI labs sought control over what outside evaluators publish, Reuters says
Anthropic and OpenAI have pushed for a big say over what outside reviewers of their AI models can make public, Reuters reported today, citing two people who know about the talks. Both companies worry that the findings could expose trade secrets. Both labs had recently promised outside reviewers a free hand. Last month, Anthropic boss Dario […] This story continues at The Next Web
Exclusive-Behind AI industry’s call for independent reviews, a fight over what gets made public
By Aditya Soni NEW YORK, Oct 13 (Reuters) - Anthropic and OpenAI have sought significant control over what outside evaluators can publish, concerned that findings could reveal trade secrets, according to two people familiar with the discussions. Anthro...
Coverage Details
Bias Distribution
- 34% of the sources lean Left, 33% of the sources are Center, 33% of the sources lean Right
Factuality
To view factuality data please Upgrade to Premium
