I will start out by admitting my biases here.
I don't buy into all of the hype around artificial intelligence as it stands now, and I certainly don't trust the results in their raw form.
Now, that may sound like a "duh" statement. Who would blindly trust the results of an AI analysis or answer to a question?
Though I'm basing this entirely on anecdotal evidence, I would say plenty of people.
That said, that doesn't mean I think AI and its practical uses are completely bunk. Obviously there's still work being done to advance it in its various uses, but progress takes time and that doesn't mean there are no practical applications in the meantime.
So that means we must all be careful in how we use it, especially the ones that are most accessible to everyone. I'm speaking specifically about the systems that are or use large language models, such as ChatGPT and Copilot.
These can be useful, but I would be wary of any answer they give. I've heard several stories about people receiving answers they know to be absolutely wrong or later find out to be wrong. When corrected, these systems admit they were incorrect.
But they do it so politely, and I think that's a large part of the problem. They present answers so kindly and confidently, plus there's the built in assumption that they're pulling from the collective wisdom on the internet that they'd be able to pull factually correct information, even about their own capabilities.
The best practice I've heard for using AI came from Lori Kiel at Pyramid Global Hospitality. On a panel at the Hotel Data Conference, she spoke about her use of AI as a personal assistant to streamline some tasks and help with some analyses. The trick is, the work she lets ChatGPT do, she already knows how to do.
As she put it, she knows the process A to Z, and she lets ChatGPT do A to L.
"The reality there is, I know if it's taking me on a journey that is not in the right direction because I ultimately know how to do that work," she said. "So that's where it is important. It should assist. It should not substitute."
I believe that's the safest strategy when using any kind of AI to assist in your work. Don't let it do something or handle a subject you don't already know because you may not recognize when it gets something wrong.
Let it help you, but like back in school, check your answers before you turn it in. A healthy of dose of skepticism would do us all some good.
You can reach me at bwroten@hotelnewsnow.com as well as on LinkedIn.
The opinions expressed in this column do not necessarily reflect the opinions of CoStar News or CoStar Group and its affiliated companies. Bloggers published on this site are given the freedom to express views that may be controversial, but our goal is to provoke thought and constructive discussion within our reader community. Please feel free to contact an editor with any questions or concern.