Google Intentionally Limiting Bard's Capabilities

Plus the first classified Senate/White House briefing on AI

Author

Nathaniel Whittemore
July 11, 2023

The AI Breakdown First Five - Tuesday July 11, 2023

Today on the First Five:

  • 5. Here Comes Killer Military AI

  • 4. China To Implement New Generative AI Rules

  • 3. Meta’s Massive Governance Experiment

  • 2. Senate Holding Classified AI Briefing

  • 1. Google Intentionally Limiting Bard AI

5. Here Comes Killer Military AI

While advanced military technology isn’t new, it’s getting a lot more attention after the release of a new Netflix documentary called “UNKNOWN: Killer Robots.” The doc argues we’re entering a (literal) arms race we can’t control. Meanwhile, Times just did a big feature about the use of Palantir’s technology in Ukraine, and the US Air Force just published a video about their use of AI.

4. China To Implement New Generative AI Rules

China’s internet regulator the Cyberspace Administration is working to develop a system that would require licenses for companies seeking to release generative AI models. The country is heading toward a refinement and implementation of draft rules that were released in April, and the CCP finds itself caught between its desire to control information on the one hand and its need to keep up with Western AI development on the other.

China to lay down AI rules with emphasis on content control - FT

The Cyberspace Administration of China aims to create a system that will force companies to gain a licence before they can release generative artificial intelligence systems, the Financial Times said on Tuesday, citing people close to regulators.

www.reuters.com/technology/china-lay-down-ai-rules-with-emphasis-content-control-ft-2023-07-11

3. Meta’s Massive Governance Experiment

One of the approaches to AI alignment is to get AI to think in more democratically-aligned ways. Anthropic is exploring something they call “Constitutional AI” and now Meta is preparing a massive experiment in deliberative democratic process around generative AI.

2. Senate Holding Classified AI Briefing

Senate Majority Leader Chuck Schumer recently announced a slate of AI briefings, including at least one classified briefing with the White House. Well, that classified briefing is happening today. Details are scant, but Schumer said the White House will share how it is 

"using and investing in AI to protect our national security and learn what our adversaries are doing in AI.... Our job as legislators is to listen to the experts and learn as much as we can so we can translate these ideas into legislative action."

1. Google Intentionally Limiting Bard AI

Google’s AI leaders have been quite vocal recently. DeepMind CEO Demis Hassabis gave a long interview with The Verge, arguing among other things that he disagrees with that leaked memo that said open source AI would outpace their models. Meanwhile, Google CEO Sundar Pichai suggested in a 60 Minutes interview that the company was intentionally holding back Bard’s capabilities for fear of deepfakes and other AI issues.

BONUS: Is this what’s going on behind the scenes of GPT-4?

The Latest AI Breakdown Episodes

Thanks for reading! How did this edition hit you?

How was today's newsletter?

Login or Subscribe to participate in polls.

Signing off from the future’s past - NLW