We are proactively implementing measures from the AI Act, even though it is not yet in force. We are establishing the necessary controls, taking into account grace periods and grandfathering. This will ensure we are fully compliant once the act comes into full effect.
What are technical and organizational measures taken while processing data with AI at Teamdash?
- Transparency to end-users - when something in our product uses AI, we make sure that the end-user knows that AI will be or has been used. This enables human oversight to override/double-check/correct/ignore anything AI-generated/processed.
- Logging - we log all LLM input and outputs for auditing and post-market monitoring. For post-market monitoring we work closely with our end-users to find any shortcomings.
- Impact analysis when building features - we design all our features with candidates' fundamental rights in mind. You'll find that we have made it extremely hard for recruiters to delegate decision-making to AI. In several (current and future) features AI is only used in assisting tasks.
- Continuous improvement based on testing and feedback. When improvement opportunities are presented, we adjust our AI functionalities to improve outputs.
- Documentation and training - we have trained our support staff to understand the topic in depth so they can advise the end users in fair & equitable AI usage.
What features is AI used for?
- CV screening on objective criteria*
- Interview transcripts and summaries*
- Candidate CV summaries*
- Removing PII from candidates for blind screening*
- Candidate similarity search*
- Candidate tagging*
- CV parsing*
- Writing assistant*
- Inclusive language checks
- Data pre-filling in job ad publishing
- Translation - landing pages, stage categories and other multilingual data