EU AI Act and US Regulators Tighten Accountability Rules for Artificial Intelligence

The European Union’s AI Act, finalized in 2024, classifies AI systems by risk level and imposes strict transparency, documentation and human‑oversight requirements on high‑risk applications such as critical infrastructure and law‑enforcement tools. The United States likewise treats AI as a product rather than a legal person, leaving an “accountability gap” over who—developer, deployer or data provider—is liable when harm occurs.

Corporate lawyers warn that the growing legal demand for detailed records clashes with AI‑generated documentation. Automated meeting assistants now create multiple digital artefacts—recordings, transcripts, prompts and drafts—that can be discoverable, inconsistent with official minutes and potentially waive privilege. Companies must now preserve these machine‑made records to satisfy duties of oversight and securities‑law due‑diligence, turning the very tools meant to streamline work into new sources of liability.