AI-generated legislation raises concerns in US House

The use of artificial intelligence tools to draft legislation is becoming more prevalent in the US House of Representatives, prompting concerns among lawmakers and staff. AI platforms such as ChatGPT and Claude are being employed by House staffers and external organizations to generate legislative text. However, this trend has led to increased workloads for the Office of Legislative Counsel (OLC), which is reportedly spending more time rectifying AI-generated bills than it would take to draft them from the ground up. The primary concerns revolve around the potential for inaccuracies, unforeseen repercussions, and the introduction of errors into the legislative process.

In a separate development, a report that formed the basis for Australia's ban on social media access for individuals under 16 years old is now under examination. Allegations have surfaced in the Australian Senate that the report contains AI-generated hallucinations within its citations. The Age Check Certification Scheme (ACCS), a UK-based organization responsible for conducting the $3.48 million trial, initially denied the use of AI in its report. However, the ACCS later acknowledged employing AI for editing purposes, though it contested the characterization of the citation errors as hallucinations. This situation has cast doubt on the reliability of the evidence used to justify the social media ban for minors.