Responsible AI Framework

TRULEO takes the ethical use of AI seriously.

We have developed our Responsible AI framework as a template for ourselves and other AI vendors in the space. Below are the tenets of our framework along with a description of their goals and evidence of how we have met them.

Transparency

GOAL

Make all invested parties - officers, police unions, and community members - aware of how TRULEO works.

EVIDENCE

  • TRULEO adapts its messaging to ensure all invested parties understand the value prop for them (see Solution at truleo.co)
  • TRULEO publishes the inner workings of its models and labels publicly for all to review

Human Oversight

GOAL

Ensure humans make decisions where discretion is involved

EVIDENCE

Explainability

GOAL

Ensure TRULEO is understandable so it can be challenged

EVIDENCE

Privacy

GOAL

Protect the privacy of both officers and civilians

EVIDENCE

Accountability

GOAL

Ensure departments are accountable and liable for their AI usage

EVIDENCE

  • TRULEO makes it easy for departments to access raw data from TRULEO analysis for a department's own goals, enabling them to take accountability for their own use cases
  • TRULEO publishes a model policy for departments to adopt as a way to responsibly use the software

Bias

GOAL

Reduce and eliminate bias in TRULEO's models

EVIDENCE

Skills

GOAL

Ensure proper training for police members to operate and understand the limits of AI

EVIDENCE

  • TRULEO conducts on-site trainings and develops training materials for departments to share physically and virtually