On October 18, the Financial Times reported that AI firm Anthropic is engaging in consultations with legal experts to establish guidelines for AI agents to negotiate on behalf of users. With the increasing prevalence of personal AI agents, Anthropic foresees a scenario where these agents can interact and negotiate with one another, acting in the best interests of their users. Two individuals with knowledge of the situation disclosed that the company is seeking counsel from law firms to examine how legal professionals can uphold client interests while ensuring equitable outcomes. According to these sources, Anthropic has been delving into the 'negotiation norms' for AI models during their interactions, considering the extent to which agents should solicit human input during negotiations and the level of vigilance they should maintain towards other AI agents. Additionally, a person close to the company shared that Anthropic has also contemplated the notion of users providing signed documents to clearly outline the tasks delegated to the agents and the rules they are required to adhere to.