<aside> ℹ️ This presentation at PAIRS 2026 Online on 17th February 2026 11:00 UTC. Registered participants will receive zoom links to join the session via e-mail.
🎙️ Thread on PAIRS Discussion Server (Discord) (register first)
</aside>
The AuDIITA workshop, now in its third year and sponsored by IEEE, has evolved into a governance model where speakers inform and provoke but the thinking is done by participants. Building on 2024's success, where discussions on AI and family violence sparked the first IEEE Industry Connection Project focused specifically on this issue, the 2025 workshop brought community service providers directly into AI governance conversations for the first time.
Participatory approaches to governance often work well in groups of 10-12 people but become harder to manage as numbers increase, diluting input to superficial "feedback" rather than critically developed ideas. Theoretical frameworks such as Arnstein's Ladder of Participation and recent work on the participatory turn in AI design (Delgado et al.) recognise that most AI projects remain stubbornly at the "consult" stage where solutions are already decided. With AuDIITA, the aim is to empower communities before AI design even starts, making them co-authors of accountability frameworks.
In October 2025, 33 participants and 12 panellists including policymakers, standards representatives, community and research organisations, and students engaged through a colour-coded dot system which encouraged conversations between different participants in each round, preventing natural clustering with familiar colleagues and creating genuine cross-pollination of perspectives. The think-pair-share format moved through individual reflection, paired discussion, and group synthesis before outputs were collated in real-time and framed as questions for the policy panel.
After two intensive think-pair-share rounds (AI and family violence, youth futures and tech addiction), participants' cognitive exhaustion signaled the depth of engagement, distinguishing this process from typical consultation exercises. The inclusion of policymakers in the synthesis stage ensured that community insights were not only heard but interrogated in a transparent, dialogic setting, reinforcing the workshop's emphasis on equal footing between technical, policy, and lived-experience expertise. Community providers identified nuanced challenges invisible to technical experts, such as how the word "honour" triggers pride versus dread depending on cultural background, a critical consideration for AI systems addressing family violence in migrant communities. They reported empowerment through their ability to define both beneficial and harmful AI applications, moving beyond reactive consultation to proactive design agency.
The workshop creates a governance pathway: community insights inform policy deliberation and real-time dialogue, which feeds into IEEE standards development—translating lived experience into institutional change. While full analysis is underway, preliminary observations demonstrate that structured participatory formats operating before policy crystallises can achieve the depth and agency typically lost when participation becomes performative rather than substantive. Full findings will be presented at PAIRS26.