Skip to content
This is NOT an official site of the City of New York. Click here for official AI reports.

NYC Office of Technology and Innovation

NYC - DTPR AI Register

AI-Assisted Forensic DNA Mixture Interpretation

What it is for

Enforcement
  • STRmix is used in forensic DNA analysis to support criminal investigations and court proceedings. It interprets DNA evidence from crime scenes to determine the probability that an individual contributed to a DNA mixture, directly informing law-enforcement decisions and judicial processes.
Risk Assessment & Triage
  • The system scores the likelihood that a specific individual contributed DNA to a forensic sample. These genotype probability scores are used by analysts to assess the strength of DNA evidence linking a suspect or victim to a crime scene.
Run by
Office of Chief Medical Examiner (OCME)
Where
No fixed location
Kept
Retained Not specified in the register entry
Shared with
Accountable organization
Your copy
You cannot see the data it holds about you. What you can do

What it is for

EnforcementRisk Assessment & Triage

STRmix is software used by the NYC Office of Chief Medical Examiner to interpret complex DNA evidence from crime scenes, including samples that contain DNA from more than one person. It uses probabilistic genotyping — a statistical method that generates millions of possible DNA profiles — to calculate the likelihood that a given person contributed to a DNA mixture. The results are used to support or challenge identifications in criminal investigations and court proceedings.

What it collects and what happens to it

Data taken in

Biometric
Identifiable data
  • Forensic DNA profiles derived from biological samples collected at crime scenes. These profiles contain genetic information capable of identifying specific individuals, including victims and suspects.
Sensitive personal information
Identifiable data
  • DNA reference profiles from identified victims and suspects of crimes, used as comparison references when interpreting DNA evidence from crime scenes.

Processing

Classification & Prediction
  • STRmix uses probabilistic genotyping — a Markov Chain Monte Carlo statistical approach — to build millions of conceptual DNA profiles and calculate weighted genotype probability distributions for mixed forensic samples. It models allelic and stutter peak heights, drop-in, and drop-out behavior.

What it does

Deciding (Analytical AI)
Human decides
  • The system produces probabilistic genotyping scores and deconvolutions of DNA mixtures. Autonomy is classified as 'Informative' — DNA analysts review the output and make the final interpretive decision for use in investigations and court testimony.

Outputs

Sensitive personal information
Pseudonymous data
  • A deconvolution of genotype probability distributions listing all accepted genotype sets and their associated weights (values between zero and one). These probability weights can be linked to a named individual when a reference profile is provided, making the output effectively identifiable in investigative context.

Run by

Office of Chief Medical Examiner (OCME)
  • The Office of Chief Medical Examiner deploys STRmix in its Department of Forensic Biology to interpret complex DNA evidence in criminal investigations. The agency has used the tool since January 2017 and performed thousands of validation tests to integrate it with their laboratory procedures.

NYC Algorithmic Tools Compliance Report — STRmix

Built by

NicheVision Forensics
  • NicheVision Forensics, LLC is the vendor supplying STRmix. The underlying software was developed by the New Zealand Crown Institute of Environmental Science and Research (ESR) and Forensic Science South Australia (FSSA). The developer assisted OCME in analyzing and interpreting data during software validation.

NYC Algorithmic Tools Compliance Report — STRmix

Kept for

Retained Not specified in the register entry
  • The register entry does not specify a retention period for STRmix outputs or the forensic DNA profiles processed by the system. Retention of forensic evidence and case files is governed by OCME policies and New York State law, which are not detailed in the source document.
  • Duration: Not specified in the register entry

Shared with

Not available to me
  • Genotype probability distributions and deconvolution outputs produced by STRmix are not directly accessible to the individuals whose DNA is analyzed. Access is controlled by OCME within the context of law-enforcement and prosecutorial processes.
Available to the accountable organization
  • STRmix outputs are available to the NYC Office of Chief Medical Examiner's Department of Forensic Biology for use in criminal investigations and court proceedings.

Stored

Stored locally
  • As a municipal forensic laboratory, OCME data is expected to be stored locally within New York City jurisdiction, consistent with municipal data governance requirements. The register entry does not explicitly confirm storage location.
  • Duration: Not specified in the register entry
How to read the colours

Can it identify you?

Anonymized data
Data about people with the link to who is broken. Stripped of identifiers, blurred, aggregated, or noised so this system can’t reasonably tie a record back to an individual.
Pseudonymous data
Each person’s data is tied to a token (hash, ID, template) that lets this system recognise the same person across events, but the token itself doesn’t reveal a name. Reidentification is possible with extra information.
Identifiable data
The data either contains a direct identifier (name, address, account name, recognisable face or voice, plate number) or carries a token this system uses to look up legal identity during processing.

Who completes the loop?

Human decides
This mode suggests; a person decides what to do next. The AI is always advisory — a human is in the loop on every decision. Example: a triage tool ranks cases for a clinician who chooses which to see first.
Human executes
This mode decides; a person carries out the result. Example: an optimizer plans the day’s trash-collection routes, and drivers run them.
Autonomous
This mode decides and acts on its own. No person reviews each decision or carries out the resulting action.

Definitions from the DTPR standard. Amber is about your data, violet about who decides. The fuller the shape and the deeper the colour, the more identifying the data or the less a person is involved.

What you can do

Ask about this system

Questions go to the NYC Office of Technology and Innovation, not the vendor.

Your rights

  • Right to Algorithmic TransparencyIndividuals whose DNA evidence is analyzed by STRmix have the right to be informed that probabilistic genotyping software was used to interpret their DNA evidence in a criminal matter. The mathematical basis of STRmix results is designed to be explainable in court, and defense counsel may challenge the methodology through standard legal proceedings. Contact the NYC Office of Chief Medical Examiner for information about how the software operates.
  • Right to ContestIndividuals subject to a criminal proceeding in which STRmix output is used as evidence have the right to contest its findings through standard legal mechanisms, including cross-examination of expert witnesses and independent forensic review. The register entry notes that results are designed to be readily explainable in court.

Risks and safeguards

  • Reputational harmErroneous probabilistic genotyping results could misidentify an innocent person as a contributor to a forensic DNA mixture, leading to wrongful prosecution or conviction.Safeguard: Autonomy is set to 'Informative' — trained DNA analysts review and interpret all outputs before they are used in investigations or court testimony. OCME performed thousands of validation tests using the software to ensure it performs correctly with their laboratory standard operating procedures and genetic analyzers. The mathematics are described as readily accessible to DNA analysts so results can be explained in court.
  • Civil liberties harmDNA evidence interpreted by probabilistic genotyping software carries significant weight in criminal proceedings and could contribute to unjust deprivation of liberty if the statistical outputs are misunderstood or over-relied upon.Safeguard: The system is positioned as informative rather than determinative; analysts are trained to interpret outputs, and the mathematical basis is designed to be explainable to courts. Validation against laboratory-specific conditions was conducted before deployment in 2017.