Cecilia Ziniti
Oct 10, 2023
ICYMI - at the TechGC AI Conference last week, several hundred general counsels got together to talk about AI. Here are my 3 takeaways and analysis.
📈 Embrace the AI opportunity
All agreed: you can't just ban using AI in your company because people already use it. Depending on the survey, 80%+ of developers report using AI for software. Grammarly, the original writing AI, has 30M daily active users. Most companies use either Microsoft or Google for core office IT which both have AI features out of the box. Upshot? Outside of highly regulated industries, policies that 'push pause' on AI won't work, so GCs are thinking of the risk of _not_ leaning in. "How AI", not "if AI". Speakers Annette Hurst and Shelley McKinley made the point clearly, with PWC data that AI will have min $16 trillion impact.
🐧 Many parallels to Open Source
When open source came out, the world took time to understand and develop business models around it, just like for AI today. In 2007 when GPLv3 (the major OSS license that's copyleft/viral) came out, I wrote my firm's client alert explaining its issues. People were uncertain about open source. There were calls to ban it, policies, advisors, new experts, and armies of consultants. Since then, open source has enabled tech's dominance and growth. Linux has become the world's single most adopted software. Github was founded, has enabled over a billion open-source commits, and sold for $7.5B. Google's a nearly $2 trillion company with almost 1B lines of open-source code in its stack. Meta is a pioneer in open source. AI's parallel to open source came up a lot. Like open source 15 years ago, it's just the beginning for AI now.
💾 All that's old is new again
The consensus in the session I led? AI doesn't really need a new _legal_ framework. US copyright law, for example, is robust and extensible. It has enabled infinite and varied contract licensing possibilities, and weathered many advances in technology from Betamax, to the internet, to digital music. For simple use of AI, it suffices to review terms of use under principles of contract law, copyright, and consumer protection. For interesting human/computer issues - lots of existing precedent. Like the monkey selfie case. A photographer painstakingly set up a camera in the jungle. He attracted animals to the setup. Amazingly, a monkey named Naruto took a selfie. The holding? No copyright for the photographer because the photograph did not have a human author as required. The upshot for AI? We might need a law or statement from the Copyright Office saying you can copyright AI-generated works. But the legal analysis itself? It's the same as before GPT, apply statutes and precedents. Then go innovate!
