EU Considers Stricter Regulations for Large AI Systems

According to a Bloomberg report, European Union representatives are negotiating a plan for additional regulations on the largest artificial intelligence (AI) systems.

Reports say that the European Commission, the European Parliament, and several EU member states are talking about the possible effects of large language models (LLMs), like Meta’s Llama 2 and OpenAI’s ChatGPT-4, and the possibility of putting more limits on them as part of the new AI Act.

The goal, according to sources close to the situation that Bloomberg cited, is to keep larger models in check while not overburdening new ventures with regulations.

The sources claim that the negotiation’s agreement on the matter is still in its infancy. The AI Act and the new proposed regulations for LLMs would be analogous to the Digital Services Act (DSA) of the European Union.

Recently, EU legislators enacted the DSA, requiring platforms and websites to secure user data and scan for illegal activities.

However, the greatest platforms on the web are subject to more stringent controls. Companies in this category, such as Alphabet and Meta, had until August 28 to revise their service practices to meet the EU’s new requirements.

It is anticipated that the EU’s AI Act will be one of the first sets of mandatory AI regulations enacted by a Western government. China has already implemented its own AI regulations, which took effect in August 2023.

Companies developing and deploying AI systems would be required to conduct risk assessments, identify AI-generated content, and be prohibited from using biometric surveillance, among other measures, under the EU’s AI regulations.

However, the legislation has not yet been enacted, and member states can still oppose any of the legislative proposals.

Since the implementation of China’s AI laws, it has been reported that over seventy new AI models have been released.