Mark Osborne discusses concerns about large language models: accuracy of results, what data they can access, and hallucinations. He references a lawyer who submitted a GPT-written brief citing fake legal precedents.
We're in the early stages of LLMs like ChatGPT and really advanced AI, but it's happening really quickly in terms of the progress that's being made. But some of the concerns that I'm hearing from our clients are really around accuracy of the results and implementation of knowledge for these smart copilots. And so your customers are really worried about how and what data the LLMs will have access to, whether or not that data feeds back other models that may or may not be under the control of the company. And the other is whether or not it's hallucinating, which has been a challenge historically from some of the LLMs. And there's even a famous court case that a lawyer submitted a brief that was written by GPT, and GPT had hallucinated precedent.
How Can You Maximize Your Business Value Before Selling? with Randy Katz Video
