The Office of Legislative Counsel (OLC) of the U.S. House of Representatives has been faced with a flood of new bills created using artificial intelligence tools. Many of these documents contain glaring errors and poor wording, making the work of lawyers much more difficult.
The Office of Legislative Counsel (OLC) of the U.S. House of Representatives has been faced with a flood of new bills created using artificial intelligence tools. Many of these documents contain glaring errors and poor wording, making the work of lawyers much more difficult.
As Politico reports , citing its own sources, a growing number of congressmen and outside groups are using generative artificial intelligence, such as ChatGPT or Claude, to write legislation. The ability of neural networks to quickly generate mountains of text means that OLC lawyers spend significantly more time proofreading and rewriting proposals. According to experts, correcting AI drafts often takes longer than writing a bill from scratch.
Errors in such documents can have serious consequences, from delayed legislation to lawsuits over misquoted laws or inaccurate legal terms. For example, AI may fail to account for the nuances of funding (tax breaks, grants, etc.) or incorrectly define the term “state,” accidentally excluding certain territories or tribal peoples from federal programs.
“It’s absolutely terrifying, and no one is thinking about it yet,” said one House staffer, speaking on condition of anonymity, noting that the use of AI is becoming more common, though they try not to talk about it publicly. The increased workload from AI means that the overburdened OLC staff can be slower to respond to requests, which in turn is pushing lawmakers to introduce untested AI bills.
