EXPOSED
The modal forensic science technician splits time between physical evidence work — processing scenes, photographing, lifting prints, packaging and logging samples, running bench instruments — and screen work that AI and automated pipelines already do well: AFIS and facial candidate ranking, DNA mixture deconvolution and probabilistic genotyping, firearms/toolmark image comparison, digital-evidence triage, and report drafting. Physical evidence handling, chain-of-custody integrity, and defending a conclusion under cross-examination are the durable core; the comparison-and-write-up middle is thinning. Court admissibility standards and lab accreditation (ANAB/ISO 17025) create a human-signature requirement, but that is a regulatory and evidentiary shield, not a licensure one, and it can erode.
Headcount grew steadily across the period.
Median pay $59,150 → $72,060 -2.5% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
+12.8% 20,700 → 23,300 on the projections basis
Growing, and only partly exposed
The BLS expects +12.8% more of these jobs by 2034, and at 56/100 the work is only partly exposed — some tasks are automatable, the core of the job is not. Nothing here is in tension.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~2,900 openings a year on average, including replacing people who leave.
CriminalistBallisticianBiometricianInvestigatorCrime AnalystCriminologistCrime SpecialistForensic AnalystFirearms ExaminerForensic ExaminerForensic ScientistEvidence SpecialistFirearms SpecialistForensic SpecialistCrime Scene ExaminerLatent Print AnalystForensic InvestigatorForensic TechnologistForensic ToxicologistLatent Print ExaminerBlood Splatter AnalystCrime Scene SpecialistFingerprint ClassifierForensic Nurse Examiner
Holding it up: embodiment . Weakest point: trust premium .
Mixed — a routine tier and a judgment tier Crawling a scene grid at 2am, swabbing a bloodstain pattern on textured drywall, and casting a tire impression in mud resist automation entirely, but the tasks that eat the other half of the week — running the sample through the CE instrument, uploading minutiae to AFIS, comparing candidate lists, and typing the standardized report language — are already substantially machine-assisted, which is why this sits at 11 rather than the 15+ a pure crime-scene role would earn.
Hands-on in uncontrolled environments You work outdoor scenes in weather, decomposition, and traffic, handle sharps and biohazards, do post-mortem prints at the morgue, and operate benchtop instruments requiring manual extraction and pipetting — 15 not 18 only because a real share of your hours are spent at a comparison scope and a keyboard in a climate-controlled lab.
Certification preferred, not legally required No state licence gates the job; what protects you is ISO 17025/ANAB accreditation, proficiency testing, and the fact that a named human analyst must sign the report and survive a Daubert challenge and cross-examination — real but institutional, and a lab can reassign that signature or narrow it to a technical reviewer, which caps this at 9.
Meaningful discretion You decide what at a scene is evidence and what is left behind, whether a mixture is interpretable, and whether a comparison is an identification, inconclusive, or exclusion — irreversible calls with liberty consequences — but SWGDE/OSAC guidelines, SOPs, stochastic thresholds, and mandatory technical review constrain those decisions far more than in an unstructured expert role, placing it at 12.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (11/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (9/20) is whether the law requires a licensed human to sign. Trust premium (9/20) is whether buyers specifically pay for a person. Judgment and accountability (12/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 30 of this occupation's 56 points (54%).
Embodiment (15/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 72/100 — SAFE.
A codified rule that an accredited examiner must personally verify, sign, and be available for cross-examination on any algorithmic comparison result — e.g. a Federal Rule of Evidence 702 amendment or state analogue requiring the proponent of AI-derived latent print/firearms/probabilistic genotyping conclusions to produce the human analyst who independently confirmed them; NIST OSAC standards being made mandatory conditions of ANAB/ISO 17025 accreditation would do the same operationally. Confrontation Clause litigation (post-Smith v. Arizona, 2024) pushing courts to require the actual analyst, not a surrogate or a tool output, is the live mechanism.
Task-mix shift as probabilistic genotyping, AFIS ranking, and toolmark image comparison consume the routine tier: the residual job becomes scene reconstruction, contamination and mixture-interpretation calls, validation studies for new algorithms, error-rate and proficiency documentation, and Daubert/Frye testimony preparation. Recognisable if lab job postings shift toward technical-leader/validation roles and analyst headcount per case falls while per-analyst testimony hours rise.
State-level licensure or certification-to-practice for forensic analysts (as Texas does through the Texas Forensic Science Commission's licensing program, and New York's crime-lab accreditation regime) extended to more states, with individual license revocation as the sanction for unsupported conclusions.
Defense-side algorithmic challenges (discovery of source code, e.g. the ongoing litigation over TrueAllele and STRmix validation) making the human examiner the named party who owns the conclusion, its uncertainty statement, and any disclosure of tool limitations under Brady — plus expanded conflict-of-interest and blind-verification requirements from OSAC that place the interpretive call on a specific accountable analyst.
Contamination and chain-of-custody rules that bar remote or robot-assisted scene processing for capital and violent-crime scenes, keeping physical collection, packaging, and continuity attestation with a person who must later testify to it. Erosion route runs the other way: patrol-officer collection kits and automated bench robotics reduce technician scene time.
The limit. Trust premium has no realistic route — the buyer is a government lab budget or a prosecutor's office, not a client choosing a human, and cost pressure runs toward fewer analysts per case. The liability gains are evidentiary rather than licensure-based in most states, so they attach to the lab's accreditation as much as to the individual, and a single appellate ruling accepting tool output with a surrogate witness reverses much of it.
| Los Angeles-Long Beach-Anaheim, CA | 1,170 | $104,410 +45% |
| New York-Newark-Jersey City, NY-NJ | 800 | $81,670 +13% |
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 660 | $85,070 +18% |
| Phoenix-Mesa-Chandler, AZ | 620 | $65,220 -9% |
| Dallas-Fort Worth-Arlington, TX | 590 | $57,180 -21% |
| Miami-Fort Lauderdale-West Palm Beach, FL | 490 | $77,820 +8% |
| Tampa-St. Petersburg-Clearwater, FL | 380 | $65,230 -9% |
| Atlanta-Sandy Springs-Roswell, GA | 370 | $63,840 -11% |
| San Jose-Sunnyvale-Santa Clara, CA | 140 | $137,280 +91% |
| Chicago-Naperville-Elgin, IL-IN | 260 | $112,790 +57% |
| San Francisco-Oakland-Fremont, CA | 330 | $109,930 +53% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 56. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.