Why AI is Both Over and Underhyped

Legal needs to deploy modern technology, but the value of AI-powered solutions can be particularly hard to make sense of. Keep reading for a practical overview of AI for legal from Cockpit Counsel guest and Dianthus founder and CTO, Rob May. 

Rob has founded an AI company, made over 70 early-stage AI investments, writes one of the world’s most popular AI newsletters, and hosts the podcast Investing in AI about its impact on business models, markets, products, and consumer behavior.

What is AI? 

Rob defines AI as “software that you can put out into the world, and without changing anything, it will adapt, learn, and get better.” Most AI companies run machine-learning models that take in a bunch of new data every day, retrain themselves at night, and run again on the software the next day. 

Huh? Here’s a legal example in the natural language processing (NLP) space. Imagine you've got 100,000 contracts and you're trying to figure out where the indemnification clauses are. Yikes. There are hundreds of pages and you don’t have time to read through all of them. Plus, you can't rely on a basic text search because people use different words and phrasing. NLP-based machine learning models can analyze data and say, I know these didn't use the exact same words, but they roughly mean the same thing. With AI-powered software, you can pull those things out.

Where to Deploy AI 

According to Rob, AI is simultaneously overhyped and underhyped. He explained, “It's really weird as a technology. The whole, artificial general intelligence killer robots thing is way overhyped. I see no evidence that we are on the track for that anytime soon. And I'm on the cutting edge of investing in this stuff. We're a long way from that. But the flip side is, it's very under-applied in most organizations.”

On this note, the feedback from Rob for business leaders is, to know where to look when it comes to applying AI at your organization. He calls his framework for evaluation the PAC framework. It stands for predict, automate, and classify. 

Those are the three common things that you can do with AI. 

• You can use it to predict things better than a human can or more efficiently than a human. 
• You can use it to automate things they used to have to do. 
• You can use it to classify things into certain buckets.

Take each department of your business and make three columns. If you’re a legal leader, go for it with legal.

• What could you predict that would help run the business better? Think, contract review times. 
• What could we automate that would help us run the business better? Think, contract workflows. 
• What could we classify that might help us run the business better? Think, contract type.  

Then go through your list. Some things are feasible and some may not be, but it’s a great place to start – especially if you’re evaluating AI-powered solutions for legal. 

Want to learn more from Rob? Watch the full episode of Cockpit Counsel. And, if you’d like to learn how we develop our AI here at LinkSquares, check out this blog post.