The Center for Artificial Intelligence and Digital Policies (CAIDP) filed a complaint alleging that GPT-4 violates FTC rules against unfair practices. A group of ethics activists in the artificial intelligence field has requested that the OpenAI startup be investigated for consumer protection law violations because the startup has introduced AI-based text generation tools that are biased, deceptive, and pose public safety risks.
Today, the Center for Artificial Intelligence and Digital Policies (CAIDP) announced its complaint after an open letter was published. The organization called for a halt to large-scale AI experiments. Mark Rotenberg, the head of CAIDP, along with several artificial intelligence researchers and Elon Musk, who was also one of the founders of OpenAI, signed the letter. CAIDP’s complaint calls for a slowdown in the development of AI models and more precise government oversight measures in this area.
The complaint CAIDP refers to potential threats of OpenAI’s text generating model, GPT-4, which was announced in mid-March. The complaint mentions solutions that GPT-4 can use to generate malicious codes and organized advertisements. CAIDP also referred to methods that biased training data can lead to stereotypical or unjustified racial and gender preferences in employment. The complaint also addresses significant shortcomings in protecting user privacy when using OpenAI products, including the ChatGPT history problem, which resulted in information about some users being shared with others.
According to The Verge, OpenAI clearly referred to the potential threats of artificial intelligence text generation, but CAIDP’s argument is that GPT-4 passes through the dangerous consumption line and as a result, more precise monitoring actions should be taken on it. The organization intends to condemn OpenAI for violating Section 5 of the FTC law, which prohibits unfair and deceptive practices. This complaint alleges that the Artificial Intelligence company has knowingly released its GPT-4 model with full knowledge of the aforementioned risks, including potential biases and harmful responses. Additionally, CAIDP considers the uncontrollable hallucinations of generative models as a type of deception. This organization says that OpenAI promotes deceptive commercial statements and advertisements with its AI tools.
CAIDP is asking the FTC to halt any commercial deployment of GPT models and to independently evaluate them prior to any possible release of other generative models in the future. Additionally, CAIDP requests that reporting tools be available to the public so they can submit their complaints via this tool. The center is also looking to establish strong regulations regarding generative AI systems.
As CAIDP notes, the FTC has expressed interest in regulating AI laws and has warned that malevolent AI systems can perform various executive actions.