According to a report by the POPVOX Foundation, Congress is ahead of some of its counterparts in using generative AI tools to support legislative processes. The House of Representatives has created an AI working group and requested internal teams to catalog AI uses. While the Senate has also established a working group, experimentation with generative AI is not as advanced. AI use in Congress has focused on tasks like writing remarks, delivering speeches, and drafting legislation. The report suggests that AI could help reduce the workload of congressional staffers by assisting with policy research and constituent outreach.
The relationship between generative AI and copyright laws could shape the global AI industry, per reporting in Politico. Court cases and the interest of the US Patent and Trademark Office have led to speculation that 2024 will be a crucial year for the industry. Different copyright laws in different countries could lead to a geographic reshuffling of AI development. Japan potentially benefits from its laws that permit AI models to train on copyrighted material without obtaining a license. Legislation is being proposed to require AI companies to disclose their training data sources to ensure copyright compliance. The outcomes of court cases and potential government intervention remain uncertain, and the uneven global landscape adds further complications.
And I also wanted to highlight a piece in the Verge. Certification programs are being offered to AI companies to demonstrate that their models do not violate copyright. Fairly Trained, founded by a former vice president of Stability AI, awards the Licensed Model certification to companies that use licensed data for training. The program has already certified nine generative AI companies.
AI will not be unregulated. I expect that existing laws will be interpreted with the help of technology. Copyright and consumer protection will be the two leveraged areas of law applied. Prepare for increased scrutiny by enhancing transparency in AI development processes, particularly in data sourcing and training methods. For providers, watch for certifications for AI models to demonstrate commitment to ethical and legal compliance or, even further, demand that from vendors.