Department of Commerce says YES to AI models but with more controls | Festina Lente - Your leading source of AI news | Turtles AI

 U.S. Department of Commerce Promotes Open-Weight AI Models but Invokes More Controls

Key Points:
- NTIA report supports adoption of open-weight generative AI models.
- Recommended government monitoring to assess associated risks.
- Proposed new capabilities for security and risk mitigation.
- Regulations under discussion in both the U.S. and Europe.

The U.S. Department of Commerce, through the National Telecommunications and Information Administration (NTIA), recently released a report expressing support for "open-weight" generative artificial intelligence models, such as Meta’s Llama 3.1. The paper emphasizes the benefits such models can bring to small companies, researchers, nonprofits and independent developers, but proposes careful monitoring to manage potential risks.

Broad Access and Benefits
The NTIA report highlights how open-weight models can democratize access to advanced AI technology, enabling even small players to compete with large technology companies. According to the NTIA, limiting access to these models could curb innovation and competitiveness, crucial aspects of economic and technological development.

Comments from the Authorities
Lina Khan, chairwoman of the Federal Trade Commission (FTC), recently said that open AI models foster competition by allowing more players to bring innovative ideas to market. Alan Davidson, assistant secretary of Commerce for Communications and Information and administrator of NTIA, reiterated the importance of monitoring risk without impeding the openness and availability of these advanced tools.

New Capabilities for Security
The NTIA report proposes that the government develop new capabilities to gather and evaluate evidence on the risks and benefits of open-weight models. These capabilities include research into the safety of AI models and the development of specific indicators to signal the need for any restrictions. The goal is to ensure that innovation continues in a safe and responsible manner.

Regulatory Context
The report comes at a time of intense regulatory activity both in the United States and abroad. California, for example, is about to pass SB 1047, a bill that would impose strict cybersecurity requirements on companies developing AI models with high computing power. In Europe, the European Union’s AI Act introduces new rules on copyright, transparency and AI use, creating a stringent regulatory environment.

Industry Reactions
Meta expressed concerns about European policies, saying they could prevent the release of future open models. Similar concerns were raised by several startups and large tech companies about the California law, which was deemed too onerous.

Conclusion
The U.S. Department of Commerce, through the NTIA, has outlined a vision that balances openness and innovation with the need to monitor and manage the risks associated with open-weight generative AI models. This approach aims to maximize the benefits of AI for a broad spectrum of users while promoting active vigilance for potential dangers.