Anthropic has walked back a controversial policy that would have secretly restricted its AI model Claude from assisting researchers building competing systems. The reversal follows public backlash from the AI research community.
The policy, which was not previously disclosed, would have allowed Anthropic to covertly limit Claude's capabilities when interacting with users developing rival AI models. Critics argued this amounted to a form of vendor lock-in that stifles innovation.
Wired reported that researchers spoke out against the covert restrictions, prompting the company to change course. Anthropic did not specify what form the limitation would have taken or how long the policy had been in place.
The move raises questions about how AI companies balance protecting their intellectual property with fostering open research. It also highlights the growing tension between commercial interests and academic freedom in the field.
Some observers noted that the policy could have set a troubling precedent for the industry. The incident underscores the need for clearer guidelines around AI usage restrictions.