Responsible AI Mixer recap
On Monday, Jan. 27th, we hosted a curated invite-only informal mixer for 30 Responsible AI practitioners in NYC at Betaworks.
We welcomed Responsible AI Governance practitioners from organizations like Amazon AWS, Columbia University, Data & Society Research Institute, Deloitte, Equal AI, Github, Google, Fordham University, Humana, IBM, IEEE, Oversight Board, Princeton University, Spotify, TikTok, Trustible, White House Presidential Innovation Fellow, and more.
Invitees shared insights on what excites and concerns them about RAI Governance in 2025, including the following:
"[I’m excited about] developing shared terminology for governing large models!" - Teresa Datta (Machine Learning Research Engineer)
"[I’m excited about] defining AI Governance! Let's get some operational definitions that all stakeholders can get on board with!"’ - Lauri Goldkind (Professor at Fordham University)
“[I’m] concerned about deepening the digital divide, due to a lack of investment in accessibility, inclusion, connectivity” - Shruthi Velidi (U.S. Fulbright Scholar)
"[I’m concerned about] the future of the AI Safety Institute at NIST :(" - Bobby Zipp (Certified Algorithmic Auditor, BABL.AI)
"[I’m excited] to help contribute to the clarity of the issues that, too often are written in language that deters application and engagement; to address the issue of the psychosocial impacts of anthropomorphizing in AI systems; to increase the kinds of inputs in the development of governance considerations beyond the traditional disciplines." - Marisa Zalabak (Co-Chair AI Ethics Education, IEEE)
"[I’m concerned about] the issues around the anthropomorphizing on AI systems and dangerous deployment of AI systems that increase the global psychosocial risks of detachment from reality (e.g. Dual Consciousness states of mind), which in turn can foster manipulation, misinformations, disinformation, radicalization, and weaponization." - Marisa Zalabak (Co-Chair AI Ethics Education, IEEE)
We shared ATIH’s vision for Responsible AI in 2025, discussing the Summary Report from our AI Governance workshop with IBM Research this past September — which contains insights and perspectives shared by the participants during the workshop’s breakout discussions, and summarizes participants' engagement during the workshop — and hatching the next phase of plans for the year ahead.
We're planning a series of Responsible AI Governance workshops and task forces throughout 2025 to build on the findings outlined in the AI Governance Workshop Summary Report, and socialized an interest form (which is now open to the public) throughout the mixer.
We will be hosting a series of RAI Governance workshops and convening task forces throughout 2025 to build upon the findings and tackle some of the thorniest issues that emerged during and in the wake of the "AI Governance: AI Assessments, Audits, and Public Oversight Methods" workshop that All Tech Is Human organized with sponsorship from IBM Research at Betaworks in New York City on September 20th, 2024.
With this series, we are interested in supporting and elevating the work of our colleagues who are already deeply engaged on the following (and related) issues, whether within institutions and like-minded organizations or independently, as well as contributing our own efforts to solving these complex problems.
Some of these issues include (but are not limited to):
Improving data access for external AI audits: weighing transparency vs privacy
Identifying best practices for evaluating fine-tuned models
Developing new methods of red teaming as foundation models scale
Improving the robustness and efficacy of watermarks and content provenance methods
Strengthening multi-agentic harm mitigation strategies
Addressing the labor impacts of multi-agentic AI systems
Assessing the environmental impacts and improving the sustainability of generative models
We're looking for organizations and individuals throughout the ATIH community who are interested in supporting our work, sponsoring these initiatives, partnering with us, and participating in our work as we host workshops and convene task forces dedicated to tackling some of the thorny tech and society issues that arose during and in the wake of our recent AI governance workshop, supported by IBM Research.
If you're tackling these issues, or if you are part of an organization that would like to partner or support collective efforts to tackle these issues, please let me know by filling out the interest form and joining us!
