Industry, Civil Society, and the Coming AI Revolution
This SXSW 2024 panel discussion covers the intersection of the tech industry’s development of transformative artificial intelligence (AI) systems and the role of civil society in understanding, critiquing and shaping these technologies for the public interest. It focused on the emerging challenges and responsibilities as AI, especially large language models and generative AI, become widely deployed products impacting society.
This panel featured a diverse group of experts who have worked at the intersection of the tech industry and civil society organizations.
- Adam Conner (Moderator) — VP of Technology Policy at Center for American Progress, previously at Facebook and Slack
- Megan Shahi — Director of Technology Policy at Center for American Progress, previously at Facebook, Instagram, Twitter
- Dave Willner — Fellow at Stanford’s Cyber Policy Center, former Head of Trust & Safety at OpenAI and roles at Airbnb, Facebook
- Katie Harbath — Chief Global Affairs Officer at Duco, previously at Facebook leading global elections work
Key Takeaways
Fundamental AI Differences
- Traditional software is deterministic, pulling from databases, while modern AI is probabilistic, generating outputs without full transparency into why
- This makes governing and controlling AI model behavior more complex compared to past technologies
Rushed Scaling
- AI companies are releasing powerful generative AI tools very rapidly to billions of users before fully understanding safety implications
- Lack of time and resources, driven by competitive pressures, leads to predictable missteps like Google’s Gemini controversy
- Moderation approaches developed for social media don’t fully translate to the unique challenges of generative AI
Politicization
- AI issues rapidly get politicized, even if no partisan intent existed behind the technical decisions
- Companies need to anticipate this scrutiny and find ways to counteract the politicization of their products/actions
Advice for Startups
- For startups in AI, the panel advised deeply engaging civil society groups early, even pre-product
- Socializing products broadly upfront and considering secondary impacts will be critical as AI permeatsfurther into society
- Understand the technical nuances of the AI system being built to provide useful input to civil society critics
Collaboration
- There are calls for much more collaboration between industry and civil society to bridge knowledge gaps
- Tech workers impacted by layoffs can leverage industry experience to work with civil society groups
- Convenings and open dialogue around “impossible trade-offs” in product decisions would be beneficial
Learn More
We curated our key takeaways with social impact startups in mind. Listen to this session in its entirety to hear the many nuanced insights shared in this hour-long conversation that we simply can’t capture in our brief article.
Explore our in-depth coverage of Social Impact Sessions & Speaker Highlights, Startup-focused Sessions, and Top AI Keynotes & Sessions at SXSW 2024.
Our StartingUpGood team believes that events and conferences are great places to learn, share ideas, and innovate. We are committed to using our innovative tech tools to share key insights and learnings from top conferences. This article uses Otter.ai to create transcripts and various LLMs to generate content summaries. All work is hand-checked for quality.
StartingUpGood supports fresh entrepreneurial approaches to social impact. FOLLOW US on social media:
Check out SDGCounting for the latest news on tracking the progress of the Sustainable Development Goals. #SDGs #GlobalGoals
