Negotiations within the European Union are underway to introduce additional restrictions for large AI models, including OpenAI’s ChatGPT-4, as part of the forthcoming AI Act.
Deliberations on AI Regulations in the EU
Reports from Bloomberg suggest that representatives from the European Commission, European Parliament, and EU member states are engaged in discussions about imposing further regulations on substantial artificial intelligence (AI) systems.
Sources close to the matter reveal that negotiators are considering the implications of large language models (LLMs) like Meta’s Llama 2 and OpenAI’s ChatGPT-4. The objective is to strike a balance between regulating these models and not overwhelming new startups with excessive rules.
While discussions are ongoing, the agreement among negotiators remains in the preliminary stages. This further emphasizes a cautious approach to the potential regulations.
Drawing Parallels with the Digital Services Act (DSA)
The proposed regulations for LLMs align with the structure of the EU’s Digital Services Act (DSA). This recently came into effect. The DSA imposes standards for user data protection and the detection of illegal activities on platforms and websites. Moreover, there will be stricter controls for major platforms like Alphabet and Meta.
Companies falling under this category had until August 28 to update their service practices in compliance with the new EU standards.
The AI Act and its associated regulations for LLMs draw inspiration from the successful implementation of the DSA. This further showcases a commitment to responsible AI practices.
Positioning the EU’s AI Act in the Global Context
The EU’s AI Act is poised to be a groundbreaking set of mandatory rules for AI established by a Western government. China has already enacted its own AI regulations, effective since August 2023.
In comparison to China’s regulatory landscape, the EU’s AI Act introduces stringent measures. These include mandatory risk assessments, AI-generated content labeling, and a complete ban on biometric surveillance.
Although the legislation has not yet been enacted, member states retain the authority to voice disagreements with any proposed regulations put forth by the parliament.
Post-Implementation Developments in China
Since the enforcement of China’s AI laws, over 70 new AI models have been reported. This further highlights the rapid evolution of the AI landscape in response to regulatory frameworks.
Read More:
India To Advocate for Global Crypto Regulations at Final FMCBG Meeting in Morocco
Standard Protocol Launches on Linea Network, Plans for DeFi Expansion