The commission's recommendations come amid increasing restrictions on US investments in Chinese tech startups and sanctions aimed at limiting China's access to advanced AI chips. Despite these measures, the report notes that as of June 2024, 50 companies in China were developing AI models, compared to a small number of large US companies. OpenAI has proposed increased government funding for AI development, citing the Manhattan Project as an example of an "iconic infrastructure project" that advanced the country. However, AI safety advocates warn of potential dangers, including the risk of AI going rogue without proper risk mitigation.
Key takeaways:
- A bipartisan congressional commission has urged for a "Manhattan Project-like" program to fund initiatives to beat China in the race to build powerful AI that surpasses human intelligence.
- The commission recommends that Congress establish a project "dedicated to racing to and acquiring an Artificial General Intelligence (AGI) capability."
- The commission also suggested giving the executive branch the power to award multiyear contracts and funding for AI, cloud, and data center firms to establish "US AGI leadership."
- There are concerns about the potential dangers that could arise from race dynamics between countries and companies in AI development, including the possibility of AI going rogue without sufficient risk mitigation.