Week 1: Welcome to the Responsible Innovation Labs "Whats Happening in AI" Update
Every week we will send interesting perspectives on AI news – covering the latest in policy, responsible AI, where AI is going, and looming big questions.
July 3, 2023
Watch/Listen:
Color's co-founder, Elad Gil, sees AI as the single biggest force available to boost global equity (ex. in education, healthcare) and misregulation as a major risk to this progress.
Conviction's GP and founder, Sarah Guo, anticipates AI co-pilots going beyond task completion into start-to-finish tools as we see increases in the context data given (ex. magic.dev, Super).
Khan Academy's CEO and founder, Sal Khan speaks on how AI could save (not destroy) education. (Microsoft's team agrees we can do amazing things with today’s tools and reflects on the future of human flourishing.)
Anduril’s CEO and co-founder, Brian Schimpf, shares his take on why companies should work with government to develop thoughtful approaches to regulation.
Read:
New partnerships announced to bring generative AI tooling into enterprise data and development platforms: Snowflake and Nvidia, Databricks and MosaicML, and MongoDB and Google Cloud.
Remaking of the web is underway. AI generated-content abounds (including the worst-case type of content on the dark web). Companies respond to their data being scrapped (ex. Reddit) and second-order impacts on Google results are seen.
Costs of large language models prompt calls and ideas to do more with less, for example: hybrid AI (with cloud and edge computing), better data inputs (ex. Microsoft’s Phi-1 test), and cost-effective methods for smaller organizations.
Re: Policy
OpenAI, Microsoft, Google, and Anthropic sent thoughts on AI accountability during open comments for the National Telecommunications and Information Administration.
Asana’s CEO and co-founder, Dustin Moskovitz, helped craft 12 AI policy ideas mentioned in recent press.
Senator Schumer discusses the role of Congress in pro-innovation US AI policy with the SAFE Innovation framework launch.
Today’s capabilities depart from the last decade (ex. DALL-E and ChatGPT vs. facial recognition and Google Translate). How we define AI in laws – whether sector-specific, cross-sector, or a mix of both – warrants attention.