The nonprofit Center for AI and Digital Policy (CAIDP) has filed a complaint with the Federal Trade Commission (FTC) alleging that Chat GPT-4 is a risk to privacy and public safety.
In the March 30, 2023 complaint, CAIDP alleges OpenAI, an artificial intelligence research laboratory based in San Francisco, California, has not created sufficient guardrails for its GPT-4 artificial intelligence product. CAIDP is an AI- and digital policy-focused watchdog organization based in Washington, D.C.
OpenAI’s press team had not immediately responded to an email requesting comment at deadline. The mailbox for OpenAI’s phone number was full and not accepting messages.
According to CAIDP’s complaint, GPT-4 is “biased, deceptive, and a risk to privacy and public safety” and OpenAI has acknowledged at least some of the dangers associated with the product, including “Disinformation and influence operations,” “Proliferation of conventional and unconventional weapons,” and “Cybersecurity.” The complaint also alleges that OpenAI has disclaimed liability for the consequences that may follow, despite OpenAI documents which warn “AI systems will have even greater potential to reinforce entire ideologies, worldviews, truths and untruths, and to cement them or lock them in, foreclosing future contestation, reflection, and improvement.”
Per the complaint, CAIDP claimed OpenAI has acknowledged “GPT-4 presents new risks due to increased capability, and we discuss some of the methods and results taken to understand and improve its safety and alignment. Though there remains much work to be done, GPT-4 represents a significant step towards broadly useful and safely deployed AI systems.” Despite this verbiage, according to the CAIDP complaint, OpenAI’s work to mitigate the AI’s behavior are “limited and remain brittle” and that the company concedes “this point to the need for anticipatory planning and governance.
Among its many assertions, CAIDP claims that despite fair and equal treatment of all consumers being central to consumer protection, “OpenAI has specifically acknowledged the risk of bias, and more precisely, ‘harmful stereotypical and demeaning associations for certain marginalized groups.’”
CAIDP further claims OpenAI documents acknowledge “We found that the model has the potential to reinforce and reproduce specific biases and worldviews, including harmful stereotypical and demeaning associations for certain marginalized groups. On the OpenAI blog, the company states, ‘While we’ve made efforts to make the model refuse inappropriate requests, it will sometimes respond to harmful instructions or exhibit biased behavior. We’re using the Moderation API to warn or block certain types of unsafe content, but we expect it to have some false negatives and positives for now.’”
The complaint goes on to present several scenarios, including ones in which AIs failed to recognize or act on potential hazards to children, facilitates corporate espionage, allows cybercriminals with limited technical skills to develop malware such as ransomware and malicious code and lowers the knowledge barriers needed to create successful cyberattacks.
Through the complaint, CAIDP leadership is seeking to have the FTC initiate an investigation that would find the commercial release of GPT-4 violates the FTC’s guidance on the use and advertising of AI products, as well as the governance norms the U.S. government has endorsed. The complaint also requests the FTC:
* Halt further commercial deployment of any GPT by OpenAI
* Establish independent assessment of GPT products prior to future deployment
* Ensure that future deployment of GPT is in alignment with FTC AI guidance
* Require constant independent assessment throughout the GPT AI’s lifecycle
* Establish a publicly accessible reporting mechanism for incidents
* Initiate a rule-making that would establish baseline standards for products in the AI market sector
A full copy of the CAIDP complaint is available here: https://www.caidp.org/cases/openai/
