Week 12: What’s Happening in AI
New US and UK proposals, AI “ghost” workers, OpenAI red teaming, Anthropic Responsible Scaling policy, talks by Bill Gurley, Ray Dalio, and more.
Image: Stable Diffusion with the prompt, “Cartoon Ghost Working”
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: New US and UK proposals, AI “ghost” workers, OpenAI red teaming, Anthropic Responsible Scaling policy, Klaviyo IPO, DALL-E 3, and talks by Bill Gurley, Ray Dalio, and more.
Policy
Americas:
Microsoft’s Brad Smith comments on Senators Hawley, Blumenthal’s proposal on licenses and White House voluntary commitments
Tech giants agree AI regulation is needed at AI Insight Summit (more here, here, and here)
Senators Klobuchar, Hawley, Coons, Collins introduce AI in elections bill (more here)
California State Senator proposes regulation of AI frontier models
Europe:
UK antitrust regulator lays out seven AI principles (more here)
The EU's Digital Services Act package is here. But how strict will enforcers be?
How Silicon Valley doomers are shaping Rishi Sunak’s AI plans
New UK government advisory service to help businesses launch AI
UK’s ‘high-risk, high-reward’ research agency starts to take shape
Asia:
Taiwan Digital Minister Audrey Tang discusses AI and democracy in New York
China and EU hold talks on AI, cross-border data flow amid renewed tensions
Responsible AI
Privacy and security:
OpenAI launches a red teaming network to make its models more robust
Confidence-building measures for AI from OpenAI Geopolitics Team and UCBerkeley lab
Workers:
Science and health:
Inceptive CEO and Transformers author Jakob Uszkoreit wants to reinvent vaccines
Elon Musk’s Neuralink is recruiting patients for its first human trial
DeepMind is using AI to pinpoint the causes of genetic disease
Chan Zuckerberg Science to build AI GPU cluster to model cell systems
Information:
Resources:
Responsible Scaling Policy (Anthropic)
Guidance for use of generative AI tools at work (Aspen Digital)
CulturaX: Dataset for LLMs in 167 languages (HuggingFace)
Fine-tune your own Llama 2 (Github Repo)
Where AI is going
Humanity and nations:
Ray Dalio on the rise and fall of nations and the changing world order at All-In Summit
Artifact co-founder Kevin Systrom doesn’t believe in AI doomerism
Computer scientist Richard Sutton on AI Succession at World AI Conference
Autonomous and dexterous:
Teaching robots new behaviors and use cases in homes (more here)
Champion-level drone racing using deep reinforcement learning
Competition:
Bill Gurley warns against regulatory capture in AI (more here)
Blockchain and consumers:
AI and blockchain, with NEAR founder and Transformers author Illia Polosukhin
Why we need to examine the motivations and effects of Latinos’ fintech adoption
Compute:
SambaNova’s new chip, claimed by CEO Rodrigo Liang to collapse costs to “1/25 of the industry-standard approach”
Huawei’s Meng Wanzhou sets AI as the next decade’s focus despite US sanctions
Releases, raises, and partnerships:
Releases: DALL-E 3 integrates with ChatGPT (OpenAI), Microsoft Copilot, Bard integrates with Google services, Anyscale Endpoints, Stable Audio (Stability AI), MADLAD-400 dataset spans 419 languages (Google)
Raises: Klaviyo’s IPO, Databricks (data, $500M), Generate:Biomedicines (biotech, $273M), Helsing (defense, $223M), Writer (text, $100M), Corti (health, $60M)
Partnerships: Roblox acquires voice moderation startup Speechly
Sign up for weekly updates in your inbox here. What's happening in AI is shaped by our Charter themes and Responsible AI work.