Technology: ChatGPT raises EU challenge to regulate AI – 03/21/2023 – Tech

The ChatGPT conversational robot has put the EU (European Union) at a crossroads in its effort to regulate the use of AI (Artificial Intelligence), as it must create a way to protect users from an ever-changing technology.

The EU has been working for more than a year on a regulation that puts order in the use of AI and, at the same time, allows users to be protected, although ChatGPT has transformed the way this technology is seen by those who use it.

Created by the American company OpenAI, ChatGPT appeared in November and quickly began to be used by users amazed with its ability to answer questions clearly, write sonnets and provide information on complicated topics.

Products powered by ChatGPT have even passed medical and legal tests designed for human students, scoring high marks.

The technology also brings risks, however, as its learning system and similar competitive models are already integrated into commercial applications.

For this reason, the EU has had to go back to square one in its effort to develop a regulatory framework, until it finds a way to effectively regulate AI.

The European Commission, the executive arm of the EU, announced in April 2021, for the first time, a plan for a set of AI rules. The European Parliament hopes to finalize an AI bill this month.

The European Commissioner for Industry, Thierry Breton, anticipated that MEPs, the Commission and the bloc’s governments are working to “clarify the rules further” on ChatGPT technology, known as general purpose AI.

real challenges

Social media users had fun trying out ChatGPT when it first launched, but the technology used is not a game.

Teachers fear students will use the feature to cheat, as ChatGPT can “write” entire essays, while lawmakers fear it will be used for misinformation purposes.

Concerns about the risks have been acknowledged by senior EU officials.

“As demonstrated by ChatGPT, AI solutions can offer great opportunities for businesses and citizens, but they can also present risks. That’s why we need a strong regulatory framework to ensure credible AI based on high-quality data,” he said. Breton.

The plan is for the European Commission, the European Council – which represents the bloc’s member states – and Parliament to discuss a final version of the AI ​​law from April.

Dragos Tudorache, the Romanian MEP leading the push to pass the AI ​​law through parliament, said ChatGPT is one of the best-known examples of general-purpose AI, although derivatives exist.

Using what is known as an “extensive language model”, ChatGPT is an example of generative AI that, operating without guidance, can create original content by analyzing prior data.

“We’re going to propose a set of rules to govern AI in general and core models in particular,” said Tudorache.

Several experts say that the risks of systems like ChatGPT have always been clear, and that the warnings have been shared with EU experts ever since work on AI regulation began.

“Our recommendation at the time was that we should also regulate AI systems that have a variety of uses,” said Kris Shrishak, a technology expert at the Irish Council for Civil Liberties.

Shrishak stressed that the effectiveness of the law will depend on the final wording, but said that the drafted text “lays a good foundation” and that “it has some mechanisms to identify new risks”.

A more pressing problem would be enforcement of regulation, warned Shrishak, and so the European Parliament is now focusing on strengthening that aspect.