AI Law Drafted by ChatGPT in Costa Rica Splits Experts

Costa Rican lawmakers asked ChatGPT to draft laws to regulate AI according to the constitution

Costa Rican politicians asked ChatGPT to create a new law for them so they could focus on the need to regulate artificial intelligence.

The chatbot was instructed by Congressmen to "think like a lawyer" and create a bill by the Constitution. The legislature was then provided the final document in its entirety.

Congresswoman Vanessa Castro led the measure's introduction and told the Thomson Reuters Foundation that the initiative had received favorable and unfavorable feedback.

ChatGPT suggested that Costa Rica establish a body to control AI systems, guided by the values of accountability, explainability, bias prevention, and human rights protection.

The law was filed in May but is currently the subject of open discussion before it is sent to the parliamentary commission for revisions and additional debate in Congress.

According to Castro, we discovered that artificial intelligence is just another tool for legislation that still requires human intervention.

Costa Rica is the ninth nation in Latin America that has recently discussed or passed legislation governing AI.

Johana Obando, a congresswoman from Costa Rica, stated that she favored AI regulation but opposed the legislation since ChatGPT merely made-up facts and provisions from the national constitution.

However, her main criticism of the proposal was that it was only a "list of good wishes" with little substance, which is a typical worry concerning AI legislation being proposed in Latin America.

According to ChatGPT, "We should regulate based on fundamental rights and international conventions," Obando added. Those rights and conventions, though, what are they? They are not mentioned in the bill.

Latin American legislators are pressing for regulation inspired by the EU's AI Act, which contains guidelines forbidding the technology's use in biometric monitoring and requiring disclosure of what AI produces information.

A bill filed in Mexico in March encourages the development of an ethical framework for the use of AI based on the protection of personal information and human rights.

However, the draught, which is currently being discussed in public forums with experts and parliamentarians, did not explain the ethical framework.

The Peruvian Congress approved the first law governing AI in the area in June, and it now only needs the president's signature to become operative.

Based on the principles of digital security and ethics, the law appoints a national body to oversee the development of AI.

Congressman Jose Cueto of Peru, who oversaw the bill's introduction, claimed it was merely a minor piece of a crucially required national cybersecurity and data protection strategy.

"The heart of the law is to… create an environment in which we can make an ethical, transparent, and sustainable use of AI," said Cueto, a cybersecurity expert and former admiral.