In the United States began an investigation into the legality of the activities of OpenAI
After numerous lawsuits against the company, regulators began to think about ChatGPT algorithms.
US Federal Trade Commission (Federal Trade Commission, FTC) launched an investigation into the activities of the company OpenAI and its artificial intelligence ChatGPT for violations of consumer protection laws, according to The Washington Post.
The claims came after a Georgia radio host last month filed a lawsuit against OpenAI . He claimed that ChatGPT mentioned him in the context of criminal activity, thereby slandering him and damaging his reputation.
Earlier, in April, the mayor of an Australian city also threatened to sue OpenAI. According to him, ChatGPT announced its involvement in a corruption scandal . OpenAI has been given 28 days to fix its AI model, but no further legal action has been reported since then.
In connection with these similar disputes, the FTC required OpenAI to provide codes for their models. The commission sent a letter to the company demanding disclosure of information about the company, marketing and training methods for AI models, risk assessment, privacy practices, integration of APIs and plugins, as well as details of data collection and use.
The document also asks for the company’s contracts with partners since 2017 and correspondence about the potential ability of AI models to “make inaccurate statements about individuals” and “reveal personal information.”
The FTC emphasized that one of the priority tasks of the commission is to control software based on artificial intelligence. At the same time, OpenAI CEO Sam Altman expressed his disappointment at how information about the FTC request became public knowledge and stressed the importance of security issues regarding AI.
Last week Company OpenAI facing another lawsuit. Fiction writers this time filed class action against OpenAI in federal court, accusing the company of copyright and privacy infringement. The authors claim that OpenAI used their books when teaching ChatGPT without obtaining proper consent, acknowledgment of authorship or compensation.
Source link
www.securitylab.ru
