RVAI Lab (formerly RegenAI)

When was the last time you felt a shift in your consciousness? ??? For me, it was last weekend at the RegenAI Conference in Ashland. What a JOY it was to be part of such a forward-thinking conversation about the future of humanity and Generative AI. I felt ALIVE listening to incredible speakers, engaging in inspiring discussions, and soaking in the many delightful surprises the organizers shared along the way. ???? It would have been easy to feel intimidated in a crowd like this — thought leaders from around the world — and once upon a time, I would have fallen into self-judgment, self-criticism, and that old feeling of not being "good enough." ?? But these days, I’m THRIVING in spaces like this! I’m finally in a place where I crave the company of others who share my sense of wonder, hope, and possibility. ???? Transformative is the word that comes to mind when I reflect on the experience. If you also attended, what word comes to mind for you as you look back a week later? ?? Chris Byrne Jay Golden Heather Stafford Thor Muller Amy Muller Ian Ingram Quinn Rose Nikki Pava Raiya Kind April Rinne Karolina Lavagnino Dondeena Bradley Pavlina McGrady, Ph.D. Jerry Michalski

Had some great conversations at the RegenAI conference in Ashland, Oregon including conf. co-coordinator Thor Muller, speakers @Brittan, Toshi Anders Hoo, Heather Stafford, Cynthia Salbato, & Jordan Plawner. You can listen to the 20 minute podcast and read my impressions about the summit at https://lnkd.in/gDjibvyh

“Trust must not only be earned but structurally guaranteed” ?? Rousing talk from Amba Kak at today’s Partnership on AI Policy Forum challenging people in many consensus-seeking spaces to actually resist a growing corporate consensus about the preeminence of generative AI in conversations about the futures we want to build. ??? Resist the notion that AI doesn’t have existing regulations, there are laws on the book that can and should be enforced ?? Regulation is not the only tool in our toolbox, so we should not stop the conversation there ?? We already have the evidence to show that the way AI is being developed is NOT in the public interest, so we should shift the burden from proving harm to actually making amends ?? The role of human rights and responsible AI is not just in calling out negative impacts, it is also a key voice in envisioning what AI can and should be

The Machine Learning for Healthcare conference?in Toronto was fantastic! Here are some of my take aways: 1/ There was almost an even split on the level of excitement and concern over the use of AI in healthcare, backed by research happening at the American Medical Association. I loved the detailed findings from a study they’ve done about physicians' sentiments towards AI (N=1081 respondents) - https://lnkd.in/gnC-gbtc 2/ A hands on red-teaming exercise demonstrated major failure modes of LLMs including GPT-4o, Gemini 1.5, Llama3_1-70B-Instruct, and Mistral-7B-Instruct-v0_3. Our team found instances of anchoring bias, poorly defined and vague answers, hallucinations, and potential fairness issues. A critical aspect of the red-teaming process was that each group had a clinician who was an active part of the process and assessed the adequacy of the LLM outputs based on their real world workflows and experience. Thank you Jim Fackler, it was great collaborating with you! 3/ Evaluating health equity across the AI lifecycle - great work from the Health AI Partnership - co-designing a framework to assess equity along five dimensions: accountability, fairness, fitness for purpose, reliability and validity, and transparency. https://lnkd.in/giXmFhW3 4/ Addressing deployment challenges is a critical aspect of integrating AI in healthcare - What happens if clinicians start believing the errors model are making? Uncertainty quantification is central for building trust in Data Science and AI. Models deteriorate if you don’t update them. Without sufficient monitoring and evaluation, errors propagate when updating a model, which leads to model deterioration. There’s a need to incorporate post-deployment feedback in real-time, while monitoring for errors and biases. 5/ We need to build a delivery science for AI - traditional ML or generative AI models are not the entire solution but rather an enabling component of broader solutions within complex systems of people, processes, and technologies. Many thanks to the directors and organizers, and thank you for so many fascinating conversations I?igo Urteaga, Christopher Khoury, Suresh Balu, Alifia Hasan, Jee Young Kim, PhD, Antoaneta V., Rezwan Ahmed, Ph.D., Jim Fackler, Michael Downs, Brian Hill, Allison McNeer, Derek Chiang, Kaden McKeen, Anirudh Gangadhar, PhD!! ?? So great to get together with part of our DLA Piper AI team Angela Zheng and Sandip Panesar ?? and huge thank you Sam Tyner-Monroe, Ph.D. and Mark Sendak for being there in spirit and making it a great experience for us! Fun thing I stumbled upon in Toronto on the last day of the conference was the opportunity to scan my iris and turn it into art (the last image below). The high resolution photo was mind blowing. I bet AI could help me learn some things about my health based on that photo! Learn more about the MLHC conference here > https://www.mlforhc.org/ All accepted papers and abstracts > https://lnkd.in/g4wzmwS4

This is an absolutely thrilling day at Planet, as we prepare (in an hour!) for the imminent launch of our next-generation hyperspectral imaging Tanager satellite, onboard a SpaceX rocket. This launch is a HUGE milestone on many different levels, and a major win for planet Earth. Tanager is a world-class 'hyperspectral' imager, which means it can sense vastly more wavelengths of light than our human eyes can see, illuminating forms of change that might otherwise go undetected. Tanager's first mission will be to directly detect otherwise invisible point-source #methane and #CO2 emissions, enabling radical #climate transparency. It will enable us to ascribe emissions to specific facilities - particular oil and gas pads, landfills, etc. and create new mechanisms of accountability, and most importantly, significnat #emissions reductions. (That will be accomplished through our partnership with Carbon Mapper, whose mission is to drive greenhouse gas emission reductions by making #methane and #co2 data accessible and actionable.) Just as excitingly, there are *incredible* use-cases for Tanager's data beyond #carbon - particularly for #biodiversity. Using airborne versions of Tanager's instruments, colleagues like Greg Asner and others have already shown they can assess biodiversity over regions of the Earth; with a fleet of these satellites, we could truly assess and monitor life on Earth at planetary scale. Tanager is important not just for what it does, but for how it came about. It's a true public-private partnership -- a co-development between Planet and NASA Jet Propulsion Laboratory, with the support and partnership many other entities, including Carbon Mapper, High Tide Foundation, Bloomberg Philanthropies, and the California Air Resources Board. It is amazing to watch this vision, borne of so many amazingly talented people finally becoming a reality. I'm particularly honored to know and work with Robbie Schingler, Jeffrey Guido, Riley Duren, Greg Asner and so many others whose persistence made today possible. Learn more about Tanager here: https://lnkd.in/g9S5Jat3 Watch the Livestream of the launch here: livestream here: https://lnkd.in/gskevq5Z

We're excited to join the panel on 'Advancing Wildfire-Resilient Architecture in a Post-Climate Change California' Thursday Sept 26! Through our work on Vibrant Planet's Land Tender, we've seen firsthand how AI and systems thinking can revolutionize wildfire management and environmental stewardship. Let's explore how we can build a more resilient future together. Sept 26, 12-1:30 PM Hosted by: Studio VARA and AIA San Francisco Eventbrite RSVP: https://buff.ly/3SOrjys Learn more about our wildfire management work: https://buff.ly/4cp4ecH #AIforGood #TechForGood #ClimateTech #RegenerativeDesign #SystemsThinking #WildfireResilience #ClimateInnovation #FutureOfArchitecture #EnvironmentalAI #ClimateAdaptation