The AI governance questions your vendor must answer
The DTA’s Responsible AI Policy is not a suggestion.
If a platform uses AI, for transcription, scheduling, shortlisting, or any other function, the vendor must demonstrate alignment with the DTA Responsible Use of AI Policy. This is a requirement for Commonwealth Agencies and an emerging expectation for state Agencies.
Four questions every vendor must answer in writing. What AI functions does the platform perform? Is AI used to score, rank, or shortlist candidates? How are AI-generated outputs reviewed before they affect selection decisions? What transparency does the platform provide to candidates about AI use?
AI that scores or shortlists candidates without human review is a different risk category from AI that transcribes or schedules. Agencies need to know which category each function falls into.