Title: US Regulators Begin Investigation on ChatGPT Amidst Allegations of Wrongdoing
The US Federal Trade Commission (FTC) has recently initiated a comprehensive investigation into OpenAI’s ChatGPT, a popular AI language model. The leaked letter reveals that ChatGPT is under scrutiny for potential harm to consumers and alleged deceptive privacy and data security practices.
In response to the investigation, OpenAI must furnish extensive documentation outlining its operational processes, the development of its Large Language Models, and elaborate on the steps taken to classify material and ensure the protection of public interests during the training of these models using online data.
Failure to adequately provide the requested information can result in severe consequences for OpenAI, including hefty fines and potentially, the complete shutdown of ChatGPT in its current form. The FTC is also actively seeking details on how OpenAI’s large language models can be retrained from scratch.
This regulatory action by the US government coincides with the filing of lawsuits by prominent US authors against OpenAI, citing copyright violations. These authors claim that their works were used without their consent, recognition, or compensation in training ChatGPT’s language abilities.
In the recently uncovered FTC letter, the regulator stated that it is investigating whether OpenAI engaged in unfair or deceptive privacy or data security practices or engaged in unfair or deceptive practices relating to risks of harm to consumers, including reputational harm.
Aside from gaining considerable popularity as a groundbreaking product, OpenAI has achieved significant financial success. In May, the company was valued at $30 billion and secured the top rank on the prestigious Disruptor 50 list. OpenAI also raised $300 million in funding the previous month.
However, OpenAI’s financial stability may be at risk due to potential claims of reputational damage, copyright infringement, and misuse of private data stemming from any findings published by the FTC. Numerous users and organizations could potentially pursue legal action against OpenAI.
Yesterday, ChannelNews reported that five authors have filed lawsuits against ChatGPT, alleging unauthorized use of their copyrighted works. These authors have also targeted Meta’s AI engine, LLaMA. It remains unclear if the FTC investigation will include other AI chatbots.
The FTC’s actions may have implications for other organizations that have utilized ChatGPT for their own purposes. Among them is Microsoft, which reportedly invested $10 billion in ChatGPT and made it readily available on its platforms.
The FTC’s letter mandates that OpenAI retain all documentary materials and produce all contractual agreements entered into since 2017. To comply with the FTC’s 49 interrogatories, OpenAI must disclose information on its officers, directors, principals, owners, financial earnings, governance models, licensing, and marketing. Furthermore, OpenAI must describe in detail all its large language model products and identify third parties who accessed them.
Of significant importance, OpenAI must provide information on how it sourced its data, including whether it was scraped, purchased, or obtained through other means. The letter specifies that if the data was derived from publicly available websites, OpenAI must disclose a list of such websites and the percentage of the data corpus obtained from each one.
OpenAI is also required to elaborate on how it reviewed and assessed the content of data used to train its large language models. It must disclose the names of all individuals involved in the process and describe the methodology for retraining the models to create new versions of them.
The FTC also demands that OpenAI detail its risk and safety assessment procedures before releasing new Large Language Models. OpenAI must clarify the measures, if any, taken to prevent the inclusion of personal information in training data.
The extensive list of questions presents a potential minefield for OpenAI, exposing the company’s negligence in handling and vetting training data if sufficient precautions were not taken.
Consequently, the FTC not only possesses the authority to impose substantial penalties but also sets the stage for protracted lawsuits, both in the United States and overseas, as multiple parties may seek legal redress.
Given the extensive investigation timeline, it is possible that Elon Musk, who had a falling-out with OpenAI CEO Sam Altman regarding the company’s direction, could become embroiled in the matter.
Ironically, Altman expressed disappointment over the leak of the FTC’s letter, emphasizing the need to build trust. He stated on Twitter, It is very disappointing to see the FTC’s request start with a leak and does not help build trust. That said, it’s super important to us that our technology is safe and pro-consumer, and we are confident we follow the law. Of course, we will work with the FTC.
