Administrative Law Judges, Adjudicators, and Hearing Officers

Right now, AI is just starting to help with judges’ routine work. For example, courts in some countries use AI to sort cases and skim documents. In Estonia, a pilot “robot judge” was designed to handle small claims by having people upload evidence and then issuing a draft decision ^[1].

In the U.S., researchers are testing AI for unemployment claims: one project has an AI model that flags which claims need more fact-checking, helping human adjudicators focus their review ^[2]. Some systems even do legal research or drafting: Brazil’s labor courts use an AI tool (“Chat-JT”) to help find cases and write standard case summaries ^[3]. Peru’s courts used an AI called “Amauta Pro” to draft protections in domestic violence cases, cutting the writing time from hours down to seconds ^[3].

These examples show AI handling the easier parts of the job – scanning documents, pulling up relevant laws, or filling out templates – while people still do the critical thinking. Tasks like swearing in witnesses, interviewing claimants, and making final legal judgments remain done by human judges. Experts note that AI should be a tool under human oversight, “part of continuous improvement,” not a full replacement for a judge’s reasoning ^{[4] [3]}.

How fast AI tools are adopted depends on many factors. On the plus side, the technology already exists for things like case searching and transcription. Governments facing heavy workloads may welcome AI support.

For example, during the COVID-19 surge U.S. unemployment offices saw claims jump by 3,000% while staffing was very low ^[2], which spurred interest in AI to speed up processing. Automating routine tasks could save time and money if implemented well ^{[2] [2]}.