US studies to apply accountability rules with ChatGPT

ChatGPT, a remarkable AI program with the ability to answer questions very quickly, caught the attention of many US congressmen as it became one of the fastest growing mass applications in history, with more than 100 million new users every month.

The National Telecommunications and Information Administration (NTIA), an agency of the U.S. Department of Commerce that advises the White House on information and communications policy, is assessing whether steps should be taken to ensure systems are secure. AI system that is “legal, effective, unethical, safe and reliable”.

“Responsible AI systems can bring enormous benefits, but only if we prevent and deal with potential consequences and harm. For such systems to reach their full potential, companies and users need to trust them,” said NTIA leader Alan Davidson.

Last week, US President Joe Biden said that it remains to be seen whether AI is dangerous. “In my view, tech companies have a responsibility to make sure their products are safe before they are released to the public,” he said.

ChatGPT surprises users with quick answers to their questions, but also annoys some by providing inaccurate information. This tool is a product of OpenAI, based in California, and invested by Microsoft Corporation.

The NTIA intends to compile a report on “efforts to ensure AI systems function as advertised and do not cause harm.” The agency said it will keep the government informed of the work it is doing to “ensure the government has a consistent and comprehensive approach to the risks and opportunities associated with AI”.

On April 11, Spain’s data protection authority asked the European Union’s privacy watchdog to assess privacy concerns related to ChatGPT.

Meanwhile, France’s privacy regulator said it was investigating numerous complaints about ChatGPT, while Italy’s data regulator was considering measures proposed by OpenAI in response to concerns that led to the rise of ChatGPT. to the Italian government temporarily banning this chatbot from March 31.