ChatGPT faces another problem as EU chief suggests new AI rules to tackle risks

European industry chief Thierry Breton said the proposed new rules on artificial intelligence will aim to address concerns about the risks associated with the ChatGPT chatbot and artificial intelligence technology, in the first comments on the application of a senior Brussels official.

Just two months after its launch, ChatGPT – which can generate articles, essays, jokes and even poetry in response to prompts – has been ranked as the fastest growing consumer app in history. .

Some experts have raised fears that the systems used by these apps could be misused for plagiarism, fraud and spreading false information, even though AI champions are hailing it as a technological leap forward.

Breton said the risks posed by ChatGPT – a brainchild of OpenAI, a private company backed by Microsoft – and AI systems underscored the urgent need for rules he proposed last year in an attempt to set the global standard for technology. The rules are currently under discussion in Brussels.

“As ChatGPT shows, AI solutions can provide great opportunities for businesses and citizens, but can also present risks. That’s why we need a strong regulatory framework to ensure reliable AI based on high-quality data,” he told Reuters in written comments. .

Microsoft declined to comment on Breton’s statement. OpenAI — whose app uses a technology called generative AI — did not immediately respond to a request for comment.

OpenAI said on its website that it aims to produce artificial intelligence that “benefits all of humanity” as it attempts to build safe and beneficial AI.

According to the draft EU rules, ChatGPT is considered a general-purpose AI system that can be used for multiple purposes, including high-risk ones such as screening candidates for jobs and rating credit.

Breton wants OpenAI to work closely with downstream developers of high-risk AI systems to enable their compliance with the proposed AI law.

“The mere fact that generative AI was recently included in the definition shows the speed at which the technology is developing and that regulators are struggling to keep up,” said a partner at a US law firm. .

“HIGH RISK” CONCERNS

Companies fear having their technology categorized as “high-risk” AI, which would lead to stricter compliance requirements and higher costs, according to executives at several companies involved in artificial intelligence development.

A survey by industry body AppliedAI showed that 51% of respondents expect their AI development activities to slow down due to the AI ​​law.

Effective AI regulation should focus on the most risky applications, Microsoft President Brad Smith wrote in a blog post on Wednesday.

“There are days when I’m optimistic and times when I’m pessimistic about how humanity will use AI,” he said.

Breton said the European Commission was working closely with the Council of the EU and the European Parliament to further clarify the rules of the AI ​​Act for general purpose AI systems.

“People should be made aware that they are dealing with a chatbot and not a human being. Transparency is also important when it comes to the risk of bias and misinformation,” he said.

Generative AI models need to be trained on a huge amount of text or images to create an appropriate response leading to copyright infringement claims.

Breton said upcoming discussions with lawmakers on AI rules would cover those aspects.

Concerns about plagiarism by students prompted some US public schools and the French university Sciences Po to ban the use of ChatGPT.

© Thomson Reuters 2023