The first government review of ChatGPT may come from the US Federal Trade Commission, OpenAI: not yet trained GPT5

produced by Big Data Digest

Following the receipt of an open letter signed by hundreds of well-known artificial intelligence experts, technology entrepreneurs and scientists, ChatGPT became the " The target of public criticism.”

On March 30, the U.S. Federal Trade Commission (FTC) received a new report from the Center for Artificial Intelligence and Digital Policy (CAIDP), requesting an investigation into OpenAI and its product GPT-4.

The report claims that the FTC has stipulated that the use of artificial intelligence should be "transparent, explainable, fair, and empirically reasonable while promoting accountability," but claims that OpenAI's ChatGPT4 "does not meet these requirements." ” and is “biased, deceptive, and poses a risk to privacy and public safety.”

CAIDP is an independent, nonprofit research organization based in Washington, D.C., that specializes in “evaluating national AI policies and practices, training AI policy leaders, and promoting democratic values ​​in AI.”

ChatGPT’s first government review may come from FTC

Legal experts say that with the growth and development speed of artificial intelligence, the FTC may formulate rules on artificial intelligence in 2023.

A December 2022 article revealed that the U.S. Federal Trade Commission (FTC) may be introducing federal regulations on artificial intelligence, although AI-focused bills filed in Congress have yet to gain significant support. ,

The article stated: “In recent years, the Federal Trade Commission has issued two publications, indicating that the focus on artificial intelligence regulation will further increase.” At the same time, the Federal Trade Commission has developed in enforcing various regulations AI expertise such as the Fair Credit Reporting Act, the Equal Credit Opportunity Act, and the FTC Act.

In addition to the regulations that have been introduced and the upcoming rules for AI, articles published by the FTC also show trends in AI supervision.

10 days ago, the U.S. Federal Trade Commission published an article titled "Chatbots, deepfakes, and voice clones: AI deception for sale" Business blog post by Michael Atleson, attorney in the FTC's Advertising Practice Unit.

The FTC Act’s “prohibition of deceptive or unfair conduct may apply to the manufacture, sale, or use of instruments that are effectively designed to deceive,” the blog post said. Even if that’s not its intent or sole purpose.” Businesses should even consider whether they should make or sell AI tools and whether they effectively reduce risk.

"If you decide to manufacture or offer such a product, you should take all reasonable precautions before placing it on the market," the blog post states. “The U.S. Federal Trade Commission (FTC) prosecutes companies that spread potentially harmful technologies without taking reasonable steps to prevent consumers from harm.”

In another article in February, “Control Your Artificial Intelligence Claims," ​​Atleson also wrote that the FTC may be "questioning" whether a company advertising an AI product is aware of the risks.

“Before you launch it on the market, you need to understand the reasonably foreseeable risks and impacts of your AI product. If something goes wrong (perhaps failing or producing biased results), you can’t just blame it. Third-party developers of this technology. You can’t say that you have no responsibility because this technology is a “black box” that you can’t understand or don’t know how to test.”

More than a hundred technology tycoons “denounced” ChatGPT, OpenAI: Not yet trained ChatGPT5

Federal scrutiny may not come soon enough, but the tech community has expressed concentrated concerns about the rapidly growing ChatGPT.

Recently, an open letter signed by hundreds of well-known artificial intelligence experts, technology entrepreneurs and scientists called for a moratorium on the development and testing of artificial intelligence technologies more powerful than the OpenAI language model GPT-4 in order to Carry out appropriate research into the possible risks.

The report warns that language models like GPT-4, which can already compete with humans in an increasing number of tasks, can be used to automate work and spread misinformation , and even raised concerns that artificial intelligence systems could replace humans and reshape civilization.

"We call on all artificial intelligence laboratories to immediately suspend the training of artificial intelligence systems more powerful than GPT-4 (including GPT-5 currently being trained) for at least 6 months." The letter reads.

Signatories of the letter include Yoshua Bengio, a professor at the University of Montreal, one of the founders of modern AI, and Jaan Tallinn, a historian and co-founder of Skype ), and the famous Elon Musk.

The letter was written by the Future of Life Institute, an organization focused on technological risks facing humanity. The letter added that the pause in research should be "public and verifiable" and should involve everyone working on advanced AI models like GPT-4.

Microsoft and Google did not respond to requests for comment on the letter. The signers appear to include employees from a number of tech companies building high-level language models, including Microsoft and Google.

In response, OpenAI spokesperson Hannah Wong said that the company spent more than six months studying the security and calibration of GPT-4 after training the model.

Hannah Wong added that OpenAI is not yet training GPT-5.

Related reports:

​​https://www.php.cn/link/9ad97add7f3d9f29cd262159d4540c96​​

The above is the detailed content of The first government review of ChatGPT may come from the US Federal Trade Commission, OpenAI: not yet trained GPT5. For more information, please follow other related articles on the PHP Chinese website!