Washington, United States. – The US Federal Trade Commission is investigating OpenAI to determine whether popular ChatGPT harms consumers by generating false information and whether its technology misuses user data.
Microsoft-backed OpenAI was told to investigate a 20-page questionnaire, which asks the company to describe incidents in which users have been falsely defamed and share the company’s efforts to ensure they don’t return.
The US regulator’s investigation was first reported by The Washington Post.
OpenAI’s launch of ChatGPT last November stunned the world by showcasing the power of Large Language Models (or LLMs), a form of artificial intelligence (AI) known as “generative AI” that can produce content similar to the content you Humans created it in seconds.
Amid the hype about the technology’s capabilities, authorities have received reports that these models can also generate offensive, false, or simply strange content, sometimes called “hallucinations.”
Commerce Committee Chair Lena Khan addressed a congressional committee hearing on Wednesday, and though she didn’t mention the investigation, she told lawmakers her agency had concerns about ChatGPT’s potentially defamatory output.
“We have heard of reports where sensitive information of people appears in response to someone else’s question,” Khan said.
“We’ve heard about defamatory statements and slander, absolutely untrue things coming. This is the kind of fraud and deceit that concerns us.”
The regulator’s investigation focuses on how this aspect can harm users, according to the survey, but also delves into OpenAI’s use of private data to build its world-leading model.
The company’s GPT-4 is the underlying technology behind ChatGPT, as well as dozens of other programs from companies that pay OpenAI to access its model for their own uses.
In a tweet, OpenAI CEO Sam Altman lamented that news about the research was leaked to the media, which “doesn’t build trust.”
Of course, he added, OpenAI will cooperate with the FTC and emphasized that “it is very important to us that our technology is safe and consumer-friendly, and we are confident in complying with the law.”
An investigation by the Commerce Commission does not necessarily imply further action, and the Commerce Commission may close the case if satisfied with the response of the examined company.
If the regulator becomes aware of illegal or unsafe practices, it will require corrective action and possibly a lawsuit.
OpenAI and the Commerce Commission did not respond to a request for comment.
