JPMorgan has told ChatGPT to clear its desk and head on home.
What happened: Employees at one of the world’s top investment banks are no longer allowed to use OpenAI’s artificial intelligence chatbot at work. And JPMorgan isn’t alone: Amazon and the tech consultancy Accenture have also banned the bot.
- While AI chatbots can be useful when writing articles or passing MBA-level exams, ChatGPT has been found to make factual errors and even try to split up marriages.
Why it matters: AI chatbots may make stuff up and behave erratically, but those banning the bots are mostly concerned they can’t keep their mouths shut. ChatGPT constantly absorbs new information to improve its language model and could repeat things to future users.
- One Amazon lawyer warned ChatGPT’s answers could “include or resemble our confidential information,” adding, “I’ve already seen instances where its output closely matches existing material.”
- There’s also a risk that sensitive information might reach the eyes of OpenAI’s human “AI trainers” who review ChatGPT’s conversations for quality control.
Bottom line: A decade ago, Wall Street tried to ban other new and disruptive tech platforms like Twitter, Facebook, and YouTube from its offices. Employees found workarounds, and the bans didn’t last. Can we really expect ChatGPT to take this latest ban lying down?