TAIPEI (Taiwan News) — ChatGPT maker, OpenAI, is being investigated by U.S. regulators over artificial intelligence (AI) risks.
According to the Financial Times on Thursday (July 13) the investigations are led by the U.S. Federal Trade Commission (FTC). Its principal mission is the enforcement of civil antitrust law and the promotion of consumer protection.
This investigation comes as Taiwan is said to be launching its own version of an AI powered ChatGPT-like chatbot by the end of 2023. Prudence would dictate that Taiwan’s chatbot developers follow the developments in the U.S. and if lessons need to be learned and changes made to the proposed Taiwan model, make them now, not after problems occur domestically.
It is understood the FTC’s investigation will include understanding, at the most basic level, how ChatGPT was designed. The investigation will also focus on ChatGPT’s so-called "hallucinations" and what oversight human reviewers have over the chatbot.
Details of consumer complaints and how OpenAI has dealt with those complaints have been requested. OpenAI has also been asked to explain how it assesses whether the consumers of ChatGPT understand issues surrounding the accuracy and reliability of the chatbot.
The issue relating to hallucinations occurs when AI, specifically a large language model (LLM) such as ChatGPT, generates factually incorrect or nonsensical information that may look plausible, such as bibliographies of names and books that don't exist and long lists of bogus legal citations. At least one U.S. attorney has already found out the hard way about the use of bogus legal citations.
Taiwan's chatbot developers must learn now how to prevent such LLM hallucinations.
'Unfair or deceptive'
The FTC is also examining whether consumers have been harmed by the chatbot’s creation of false information about them, as well as whether OpenAI has engaged in “unfair or deceptive” privacy and data security practices. This is in accordance with the FTC’s mandate of consumer protection.
Similarly, Taiwanese consumers warrant appropriate protection both domestically and internationally.
Generative AI products, like ChatGTP, have been under close watch by global regulators. Much of the concern relates to the enormous amount of personal data accumulated by the technology. If that data is used in ways that breach personal data protection laws, then chatbots face considerable ongoing legal issues.
The generative AI industry cannot claim to be shocked by the investigation. In May, the FTC clearly signaled its concerns and possible intentions, perhaps hoping for some sort of "house cleaning" by the industry.
The U.S. regulator has asked OpenAI to share internal material ranging from how the group retains user information to steps the company has taken to address the risk of its model producing statements that are false, misleading, or disparaging. The request — and only the ill-advised or appallingly arrogant would not see this as a demand from a major regulator — means that OpenAI is now facing a very broad investigation.
The FTC has alleged that it has received reports of people’s sensitive information showing up in response to an inquiry from a third party. The regulator also alleges libelous, defamatory statements, and flatly untrue facts are emerging as a result of searches.
Such results are potentially fraudulent and that leads to a lengthy legal highway. The FTC has said that its primary concern is that Open AI is following the law.
Watch and learn
The regulator is not making noises about shutting down the technology, simply ensuring the technology is within the boundaries of existing law. Changes in the law may be necessary when the results of the investigation are published and as the chatbot industry evolves, and this applies as much to Taiwan as to the U.S.
The FTC, by launching this investigation, must believe it has the resources to complete what it has now started.
OpenAI had more than 100 million monthly active users two months after its launch, in November 2022, and even those numbers alone indicate that OpenAI and the FTC will be reviewing staggering amounts of data. Does the Taiwan regulator have the resources, knowledge and expertise to launch a similar investigation should that be warranted?
No doubt expecting the announced investigation, OpenAI has made its privacy policy more accessible and introduced a tool to verify users’ ages, but whether this move and the answers it ultimately provides to the FTC under the terms of the investigation are sufficient remains to be seen.
Let us watch and learn from the FTC's investigation and ensure Taiwan does not make any similar mistakes.
