Forensic Scientist

A lot of forensic lab work is pattern matching, and AI is good at pattern matching. Fingerprint systems search databases of millions of prints in seconds and hand the examiner a ranked shortlist. At the forensic institute of France's national gendarmerie (IRCGN), most routine, good-quality DNA profiles are now processed automatically, with a technician spot-checking a random sample of cases each week. A January 2026 review of 61 studies on AI fingerprint identification (Forensic Sciences, MDPI) found AI speeds up searches and can help cut backlogs, though evidence on cost savings was limited. But it also found that examiners still make errors, that people tend to trust the machine too much, and that courts need AI results to be explained and validated. Its conclusion: AI works best as a first filter or a decision aid, with "the final decision ... always made by a trained expert."

That is the rubric's picture of a 2: AI takes over parts of the work, but demand absorbs the time saved. The Bureau of Labor Statistics projects forensic science technician jobs to grow 13% from 2025 to 2035, much faster than average, with about 2,800 openings a year. BLS expects labs to keep hiring to handle "high caseloads" and says better science and technology will make forensic evidence more useful in court, which increases demand for the people who produce it. Our read is that working crime scenes and defending results under cross-examination also stay with people. The field is small, though: about 20,100 jobs.

We moved this from 3 to 2. The old score rested on a "35% of labs use AI" figure from a careers website that named no study, and on a claim that labs need fewer analysts per case, which no source supported. What would move it back to 3: labs automating routine DNA and fingerprint work far enough to hire fewer entry-level analysts, or BLS cutting its growth projection.