Synthesis of 161 Empirical Studies on Responsible AI Practices in Industry
A thorough review of literature encompassing 161 empirical studies conducted over six years consolidates existing insights into Responsible AI (RAI) practices within various industries. Engaging industry professionals through interviews, surveys, workshops, and ethnographies, the studies indicate significant advancements alongside ongoing obstacles. While awareness among practitioners has grown, RAI initiatives have become more structured, and the use of toolkits and guidelines has increased. Nonetheless, implementing and maintaining RAI efforts remains challenging for many. The review underscores the variability of findings across different roles, organizational settings, and interventions, aiming to unify this knowledge. This research is available on arXiv with the identifier 2608.10431, categorized as 'cross'.
Key facts
- The review synthesizes 161 empirical studies on Responsible AI practices in industry.
- The studies span six years and involve industry practitioners.
- Methods used include interviews, surveys, workshops, and ethnographies.
- Practitioner awareness of RAI has increased.
- RAI activities have become more professionalized.
- Interventions such as toolkits and guidelines are more widely adopted.
- Persistent challenges remain in implementing and sustaining RAI work.
- The paper is available on arXiv with identifier 2608.10431.
Entities
Institutions
- arXiv