What should I look for when evaluating AI supplier tenders?
Answered
Use a multidisciplinary evaluation team. Look for ethics approaches, accountability for algorithm outputs, bias mitigation, reproducibility, testing under varied conditions, defined performance, security, and the skills and diversity of the delivery team. Expect approaches to evolve — AI delivery is iterative.
From the guidance
Primary (how) Guidelines for AI procurement
Section: 3. Selection, Evaluation and Award
Read this in Guidelines for AI procurement (opens in new tab)
Secondary (normative) Guidelines for AI procurement
Section: 9. Focus on the need to address technical and ethical limitations of AI deployment during your evaluation
Read this in Guidelines for AI procurement (opens in new tab)
Related questions
- How do I build equality and data protection into AI tenders?
- How should commercial teams apply Secure by Design when buying AI?
- Who is responsible for security when we use third-party AI components?
- How do I avoid black-box algorithms and vendor lock-in when buying AI?
- How do I specify requirements when buying AI?