Utilising the Assessment List for Trustworthy AI: Three Areas of Improvement
Gavornik, A, Podrouzek, J., Mesarcik, M., Solarova, S., Oresko, S., Bielikova, M.
This position paper contributes to the discussion on the utilisation of Assessment list for Trustworthy AI (ALTAI). Building on our own empirical experience stemming from cooperation with various AI teams and organisations, we suggest that ALTAI is suitable to be used as the core of the ethics-based assessment process. However, there are several areas in which ALTAI needs to be supplemented by additional tools and approaches. We introduce three extensions of ALTAI: (1) stakeholder analysis and segmentation; (2) data flows mapping; and (3) management of risks and countermeasures. We suggest that such enrichment will help ALTAI to narrow the gap between abstract principles and requirements of trustworthy AI and to increase the chances of implementing more ethical AI systems. Moreover, such an approach to ALTAI is also compatible with requirements expressed by existing and proposed EU legislation.
Cite: Gavornik, A, Podrouzek, J., Mesarcik, M., Solarova, S., Oresko, S., Bielikova, M., Utilising the Assessment List for Trustworthy AI: Three Areas of Improvement. Proc. of the 2nd Workshop on Adverse Impacts and Collateral Effects of Artificial Intelligence Technologies – AIofAI 2022, Vienna, Austria, July 24, 2022, CEUR-WS.org, online https://ceur-ws.org/Vol-3275/paper3.pdf