GPT-4, the new language model with which OpenAI replaced GPT-3.5, could be the company's last release for a long time. Tech ethics group Center for Artificial Intelligence and Digital Policy (CAID) is asking the US Federal Trade Commission to block OpenAI from issuing new commercial releases of the chatbot that has captivated some users and caused distress to others with their quick, human-like responses to queries.
This organization filed a complaint with the US regulator on Thursday, arguing that GPT-4 is "biased, misleading and poses a risk to privacy and public safety", in addition to violating "federal consumer protection law". .
CAIDP asked the FTC to "open an investigation and then suspend further deployment of commercial GPT products until the company complies with the FTC Guidance for Artificial Intelligence (AI) Products," the technology ethics group explains in a statement. release.
“The FTC has a clear responsibility to investigate and prohibit unfair and deceptive business practices. We think the FTC should look closely at OpenAI and GPT-4," said Marc Rotenberg, the center's president and general counsel.
Also, remember that American AI companies have an obligation to inform and warn users of the risks of their products. So it is not enough to “simply warn your customers about misuse or tell them to disclose isn't enough to deter the bad guys”.
“Your deterrence measures should be durable, built-in features and not bug fixes or optional features that third parties can undermine through modification or removal,” they argue.
For example, OpenAI exposed private chat histories to other users, and one AI researcher found that it was possible to "take over someone's account, view their chat history, and access their billing information without them realizing it," explains the group in its complaint.
Specifically, it lists nearly a dozen major risks of OpenAI's GPT-4 system: "Disinformation and Influence Operations," "Proliferation of Conventional and Non-Conventional Weapons," and "Cybersecurity," among other issues.
OpenAI itself has recognized that "AI systems will have even greater potential to reinforce ideologies, world views, entire truths and falsehoods, and cement or block them, preventing future contestations, reflections and improvements," continues this organization.
“We are at a critical moment in the evolution of AI products. We recognize opportunities and support research. But without the necessary safeguards in place to limit bias and deception, there is a serious risk to businesses, consumers, and public safety. The FTC is in a unique position to address this challenge,” concludes CAIDP President and Director of Research Merve Hickok.
