Great webinar put together by Giulio Coppi and Access Now team, "Not AI for good", a #civilsociety-led discussion that cut through much of the optimism surrounding the AI for Good summit and adjacent meetings that happened last week in Geneva.
What made the conversation so valuable was its refusal to stay at the level of abstract principles. It asked the questions that are often left out of mainstream #AI debates:
🔹What does it mean when AI is used in #warfare?
🔹What happens when #humanitarian organisations become operationally dependent on technology companies implicated in human rights abuses, violence, or armed conflict?
🔹And how do we preserve #accountability when the systems shaping public-interest decisions are designed, owned, and updated by a very small number of private actors?
I also recently contributed to a International Civil Society Centre webinar on “critical washing” and the relations between the #privatesector and #CSOs, which reinforced a similar point: partnership language can sometimes obscure power imbalances, conflicts of interest, and accountability gaps unless it is matched by real scrutiny and safeguards. (cc Rachel Wilkinson Elizabeth Parsons Kavisha Pillay)
The strongest takeaway for me was that AI governance in these contexts cannot be reduced to ethics statements or voluntary commitments. If public institutions, humanitarian actors, and international organisations are going to use these tools responsibly, they need governance that is concrete and enforceable: rigorous procurement standards, human rights due diligence, contractual safeguards, transparency over model use and data flows, independent oversight, auditability, and clear exit strategies when risks become unacceptable. Without that, “innovation” can quickly become dependency.
A few points stayed with me:
🔸In humanitarian action, reliance on dominant tech providers is not just a procurement issue; it is a digital resilience issue.
🔸In security and conflict settings, the margin for error is too small to rely on trust alone.
🔸Civil society often sees the harms first, but is still too often excluded from the decisions that shape deployment.
🔸Responsible AI has to be built into institutions and contracts, not just discussed in principles.
I appreciated the perspectives brought by Tina Indrani Mason, Abeba Birhane and Luc Dockendorf. The discussion was strongest when it moved from critique to governance: not simply naming the risks, but asking what accountable, rights-respecting, and resilient use of AI should actually look like in practice.
If we want a serious conversation about AI, we need to be willing to talk about power, dependency, and harm—not only innovation and efficiency.
#AI #AIGovernance #HumanRights #DigitalRights #ResponsibleAI #HumanitarianAction #TechnologyGovernance #AIandSecurity
Missed Not AI for Good? 🚨
Catch the conversation on Big Tech's role in the humanitarian tech space, the militarization of AI and cloud systems, and the questions missing from the AI for Good Summit.
Watch the recording: https://lnkd.in/eDxTq8yw
