Week 15: What’s Happening in AI
Israel’s tech sector copes with terrorist attacks. Resources by Anthropic, DeepMind, GitHub, and more. California’s VC diversity reporting rule.
Welcome to this week’s edition of What's Happening in AI, covering the latest in policy, responsible AI, and where AI is going for the Responsible Innovation Labs community.
Below: California’s new VC diversity reporting law, EU’s DSA enforcement, and G7’s AI plans. Anthropic, DeepMind, GitHub on product and Cloudflare, AWS on security. New releases by ElevenLabs, more. Technology was supposed to prevent the Hamas attack – what happened?
Policy
Americas:
California Governor Newsom passes law mandating VCs release investments’ diversity information (more here)
Sen. Klobuchar, Rep. Clarke send Meta, X letters on AI-generated political ads (more here)
US reportedly warns EU’s landmark AI policy will only benefit big tech
Europe:
EU warns X and Meta over disinfo targeting Israel-Hamas war, re: DSA
Snap’s AI chatbot draws scrutiny from UK regulator over kids’ privacy concerns
EU sees 'convergence' with Japan on AI at UN Internet Governance Forum
More:
Southeast Asia mulls more business-friendly approach to AI rules (more here)
G7 to draw up AI code of conduct this autumn, reported to include watermarking
Responsible AI
Product and alignment:
Anthropic on understanding language models
Google DeepMind, Anthropic on AI “constitutions”
GitHub’s Mario Rodriguez on building a $100M ARR product with responsible AI
Identity and safety:
RunwayML on mitigating stereotypical biases in text to image generative systems
Twitch adds anti-harassment features to stop banned users from watching streams
Security and cryptography:
Cloudflare on HTTP/2 Rapid Reset: deconstructing a record-breaking attack
Google on passwordless by default: making the switch to passkeys
Bitcoin and Tether funnel millions of dollars to terror groups (more here)
Sciences and health:
General Catalyst wants to buy a health system as testing ground for new venture
Early exploration of OpenAI’s GPT-4V reading x-rays
Munich startup Orbem banked $32M to scan eggs, says humans are next
Microsoft announces new AI tools to help doctors deliver better care at HLTH 2023
Information and value exchange:
Medium CEO Tony Stubblebine on why the platform is opting out of AI
A doctored Biden video is a test case for Facebook's deepfake policies
Researchers, activists try to get ahead of AI-driven election misinformation
Resources:
Elections best practices guide (Integrity Institute)
Securing the digital future (Semafor)
Where AI is going
Society and technology:
Technology was supposed to prevent the Hamas attack. What happened?
Databricks' Ali Ghodsi talks about enterprise AI with a16z’s Ben Horowitz
Capabilities:
Theory’s Tom Tunguz on centaurs and cyborgs: the jagged frontier of AI (more here)
I’m totally blind. AI is helping me rediscover the world (more here, here)
Blending training audio/visual for multimodal LLMs is investigated in new ‘FAVOR’ paper
Reducing computational overhead of diffusion models is discussed in new paper
Compute:
The future of Moore's Law is inside this Willy Wonka machine
Tired of shortages, OpenAI considers making its own AI chips and Microsoft may debut its first AI chip to mitigate cost (more here)
Climate:
Releases and raises:
Releases: ElevenLabs’ AI Dubbing, Character.AI’s Group Chat, Cohere’s Coral, Adobe’s Firefly 2, Replit’s new AI model, Nucleus AI’s 22B Transformer LLM
Raises: Mana.bio ($19.5, bio), Cleanlab ($25M, data stack), Anysphere ($8M, IDE), Gutsy ($51M, security), Observe ($50M, data stack)
Partnerships:
Sign up for weekly updates in your inbox here. What's happening in AI is shaped by our Charter themes and Responsible AI work.