GPT-4, ChatGPT’s bigger brother, is just out and has already created a new market: One click lawsuits. Yes, you read correctly. How clever. Just what the world needed.
As reported by Financial Times a few days ago, the company DoNotPay, a RoboLawyer (!) company specializing in contesting parking tickets, is working on integrating GPT-4 into an app to create one-click lawsuits. Not general lawsuits yet, only lawsuits against nuisance robocallers. But we don’t have to be very creative to see where it’s heading.
It may sound like a joke, but read the preceding paragraph again and you’ll probably see both irony and insanity. The world’s most advanced software technology – a LLM in this case, which people and media increasingly refer to as ‘AI’ – is being used to fight robocalls. Essentially – robot against robot. RoboLawyers, robocalls, robo-litigation. Exactly what we need to save the world, right?
The irony (or insanity) doesn’t stop there. The LLMs, Large Language Models, have recently become (in)famous for inaccuracy, creativity presented as fact. GPT-4 is more capable than its predecessors, but doesn’t even pretend to be more truthful. Is this our new lawyer? It sounds pretty much like the old ones – as narrated in the Suits TV show, except the new one can handle a million concurrent cases while the old one can manage a couple. Incredible improvement, isn’t it?
Apparently, (part of) the next level of societal development is machine litigation. Robots suing robots while we’re watching – and paying. Seriously, if we stick to the (in)famous Trump dogma of ‘let’s see what happens’, that’s exactly where we’re heading. If we want a different direction, we need to do something about it. Now.
Will GPT-4 send lawyers out of business big time? Not a chance! Like I told a worried lawyer-friend the other day: “You have a lot of good business coming if you stay focused on real life, real people, real problems and reliable tools.” The latter ironically includes LLM-machines like ChatGPT. Not the freewheeling variant we’ve seen thus far, but a controlled and predictable tool.
Related posts on mindset3.org: