CAIDP files FTC complaint against OpenAI's GPT-4, citing risks

The Center for Artificial Intelligence and Digital Policy (CAIDP)

has taken a stand against the release of powerful AI systems to the public by complaining to the United States Federal Trade Commission (FTC). 

Specifically, the complaint targets OpenAI's latest large language model, GPT-4, which the independent non-profit research organization claims is "biased, deceptive, and poses a threat to privacy and public safety."According to CAIDP, the release of GPT-4 to the market breaches Section 5 of the FTC Act, which prohibits deceptive or unjust practices in commerce.

To support its argument, CAIDP cited the GPT-4 System Card, which revealed that AI systems can strengthen particular ideologies, beliefs, truths, and falsehoods, and solidify them, making them difficult to challenge or modify in the future. The organization emphasized the potential dangers associated with this feature, as it could limit critical reflection and improvements. 

CAIDP further claimed that OpenAI had launched GPT-4 for commercial use with full awareness of the associated risks and that no independent evaluation of the system was carried out before its release. Consequently, the organization urged the FTC to investigate the products of OpenAI and other entities operating robust AI systems.

"It is time for the FTC to act [...] CAIDP urges the FTC to open an investigation into OpenAI, enjoin further commercial releases of GPT-4, and ensure the establishment of necessary guardrails to protect consumers, businesses, and the commercial marketplace," the organization said.

The news comes at a time when concerns are growing about the development of advanced AI systems. A petition introduced by the Future of Life Institute on March 22 called for a "pause" in the development of AI systems more potent than GPT-4, and garnered support from several high-profile individuals, including Elon Musk and Apple's Steve Wozniak, as well as 2600 AI experts. 

In the petition, the authors emphasized that the advancement of AI could have a profound impact on the history of life on Earth, for better or for worse. Additionally, UNESCO has urged states to adopt the UN's "Recommendation on the Ethics of AI" framework.

GPT-4, launched on March 14, succeeded ChatGPT-3, which debuted in November. The latest model is regarded as ten times more intelligent, and a recent study revealed that it outperformed even the most challenging U.S. high school and law exams, achieving scores in the top 90th percentile.