The Collective Intelligence Project

CIP is living up to its name: we’re building tools for collective intelligence. Here’s a bit more on what we’re working on: https://lnkd.in/eG4BgaRk 1. Scenario Construction. This tool will help policymakers generate scenarios, stress-test policies, and gather public input in high-uncertainty settings. Useful for collective deliberation on complex, uncertain topics like AI governance. 2. Open-Source Collective Constitutional AI (OSCCAI). Expanding on our CCAI project with @AnthropicAI, we’re building an open-source platform that allows for communities around the world to create their own constitutions and fine-tune models. 3. Collective Intelligence with Agents (CIwA) Imagine AI agents representing different areas of expertise, deliberating on complex topics on behalf of you. Anticipating a more agentic world, we are setting up a framework for the collective governance of deliberative agents. 4. Voice of Nature: AI agents advocating for natural entities (e.g., rainforests, rivers) in human discussions. By combining environmental data with LLMs, these agents give voice to non-human stakeholders in critical conversations. These tools are early experiments in our mission to enable better collective input into important decisions. Interested in contributing? In addition to hiring for a founding engineer, we’re also looking for project partners and funders to help us build these tools for collective intelligence. Reach out!

💡Testing AI safety with “human interaction evaluations" - a novel way to explore AI safety and risks 🤓 👉 human machine interaction is a key element of applying LLMs in practice 👉 current evaluation methods are not taking that fully into account 👉 developing a new approach leads to more specific identification of additional risk factors 👉 proposal for a human-LLM interaction taxonomy. 👉 Research Paper: Beyond Static AI Evaluations: Advancing Human Interaction Evaluations for LLM Harms and Risks 👉 by Lujain Ibrahim Saffron Huang Lama Ahmad Markus Anderljung👏 👉 from Oxford Internet Institute, University of Oxford Centre for the Governance of AI (GovAI) The Collective Intelligence Project OpenAI #artificialintelligence #riskmanagement #innovation

The thing you've all been hearing me talk about for so long is real!! This is a long thank you post for finishing up my fellowship project with Open Future Foundation, and I have many people to send love to. We launched our Report on AI and the Commons last week at the Public Knowledge Emerging Technology conference! Super thankful to Nicholas P. Garcia and Christopher Lewis for making that happen. Nick was an excellent thought partner in figuring out how to meaningfully present the outcomes of our report, and I truly appreciate the trust in giving us the final speaking slot of the conference! At the conference, I spoke about our report and then had the honor of asking questions to panelists Aviya Skowron and Alek Tarkowski! Alek has been my mentor throughout this fellowship, and this collaboration has been super fun and truly wonderful. I met Alek at the Creative Commons summit last year, and I remember feeling both challenged and excited by the questions he posed. It's been an awesome few months and the Monday 11pm pst/8am cest calls were so worth it. Aviya was full of hot takes this whole week, and they really blew me away with their insights. EleutherAI is probably one of the coolest projects you'll see, and this org is truly doing the most for open source artificial intelligence. My gratitude and admiration to everyone at Open Future for their support, especially Alicja Peszkowska for all your help marketing and organizing events and Paul Keller for your discerning academic gaze! To Viviana Rangel and Maria Jose Parra Triviño of Fundación Karisma for your support with the Latin American groups! To Divya Siddarth, Saffron Huang, and Flynn Devine at The Collective Intelligence Project for the incredible support and trust! To Paula of RadicalxChange Foundation and Liz at The Computational Democracy Project for the guidance and insights! To patcon for helping us set up the Pol.is instance. To Anna Tumadottir, Angela Oduor Lungati, Joshua Tan, Derechos Digitales, COMMUNIA, Creative Commons, and Open Knowledge Foundation for speaking, supporting, and lifting this work up. To Catherine Stihler who brought me in with open arms, and to Morgridge Family Foundation for funding the Summit grant that started my project with CC! 💛

Hi All 👋 AI Tamil Nadu (Previously AI Coimbatore) is back with another exciting in-person event in collaboration with Google for Developers and TFUG Coimbatore coming Saturday! We have some amazing sessions lined up where I'll also be talking about some of the community initiatives that AITN is doing in the Tamil LLM space in collaboration with The Collective Intelligence Project ❤️ If you're interested to participate and contribute to the initiative during the event, please do register and book your seats! Seats are filling fast and it is limited 🚀 Looking forward to see you all! #opensource #llm #tamil #generativeai Divya Siddarth Saffron Huang Evan Hadfield

The amazing Jennifer Ding came to our London event two weeks ago to share her thoughts on the role of open source for democratizing AI (something we talk about a lot in our roadmap…) Main takeaway? WHO creates this technology is just as important as WHAT comes out of it. People should have a say in shaping the technology that affects their lives and the AI community must devote resources to create more pathways for people to participate in the AI pipeline. We agree! Check out our roadmap here: https://lnkd.in/gVcfkbrf 

We were so lucky to have one of today's leading AI journalists, Billy Perrigo, come to our London event last week. His message on democratizing AI? The wealth and power accrued from this tech are currently in the hands of the few; for truly democratic AI, we must fundamentally think about how ownership can be more distributed. Well put! You can see our take on this and more in our roadmap: https://lnkd.in/gVcfkbrf

Amazed by the number of people who came to our webinar yesterday! Thank you all so much for joining us. We're incredibly grateful to our speakers, Yoshua Bengio, Bruce Schneier, Amba Kak, Manu Chopra, and Teddy Lee, who all brought such fascinating insights on what it means to democratize AI. The recording is coming soon for those who coulnd't make it! But in the meantime, be sure to check out our roadmap: https://lnkd.in/gVcfkbrf

Haven't signed up to our webinar yet? Well maybe this will persuade you. We have an incredible line-up of speakers, starting with one of the most important people in the field of AI... Speaker 1: Yoshua Bengio We couldn't be more excited to have an AI legend come to speak about democratizing AI. After leading the way on deep learning, he has a better understanding of AI’s potential than most, coupled with a clear-eyed view of the risks. Speaker 2: Amba Kak Amba joins us to speak about her work as Executive Director of the AI Now Institute and her fifteen years spent designing and advocating for technology policy in the public interest all around the world. Speaker 3: Nabiha Syed Nabiha is an award-winning journalist and CEO of The Markup. She brings a wealth of experience in understanding the role of the media in societal shifts after advising and reporting on some of the most important stories of the 21st century. Speaker 4: Teddy Lee Teddy co-leads the new Collective Alignment team at OpenAI and joins us to talk about their explorations, experience running a grant scheme on the topic, and plans for the future. Speaker 5: Manu Chopra After featuring on the cover of TIME for his work leading Karya, Manu joins us to share his experience empowering Indian communities in the age of generative AI and lessons learnt for communities world over. Sign up here: https://lu.ma/8ng4lz7k

Today we’re publishing our ‘Roadmap to a Democratic AI’! You can read it here: https://lnkd.in/gVcfkbrf This is our take on what the ecosystem can build, research, advocate for, and fund in 2024 to democratize AI. Saying we want more democratic tech is one thing; doing it is another. Here's our first guess at what needs to get done... 1. Advance collective fine-tuning of frontier models - Building on our work with Anthropic around Collective Constitutional AI, this works on proven methods and is a gateway to more public input in AI. Ship public input to prod! 2. Identify other opportunities for ongoing public input in the AI lifecycle - There are multiple leverage points to incorporate ongoing public input. Let's continue to develop better methods to integrate input during development, deployment, and post-deployment of AI. 3. Connect the open-source and democracy movements - Efforts to increase access (open-source) and efforts to increase participation (democracy) tend to attract different communities; we believe the two worlds need each other. Democratization is more than access; it’s governance. 4. Keep expanding #AlignmentAssemblies and other processes to more parts of the world and more languages - Public input doesn’t do much if we have tiny publics. We need to bring in and emphasize international voices. 5. Build AI-enabled tools for democratic governance - The future of democracy should be better than the past. Building transformative tech into governance ensures collective intelligence maintains pace with artificial intelligence. 6. Experiment with institutional governance models for AI development - Collecting all the input in the world doesn’t matter if our institutions aren’t set up to care about or act on it. We need new models! Democratizing AI isn't easy and we welcome critiques and collaborations! So reach out. Want to hear more? Make sure you sign up to our webinar happening next week: https://lu.ma/8ng4lz7k Speakers include Yoshua Bengio, Nabiha Syed, Amba Kak, and Manu Chopra (more announced soon).

Want to learn more about CIP's vision for how AI and Democracy can help fix each other? Well our Executive Director, Divya Siddarth, just gave a TED Conferences talk exploring exactly this! The main takeaway? As Divya puts it, "We are told that transformative technologies like AI are too complicated, or too risky or too important to be governed democratically. But this is precisely why they must be. If existing democracy is unequal to the task, our job is not to give up on it. Our job is to evolve it. And to use technology as an asset to help us do so." Watch the whole talk here to find out more: https://lnkd.in/g5s2qdKq