With the continued rise of AI generated content, misinformation and disinformation continue to proliferate social media and online spaces. At the same time, more and more people are putting their trust in AI. Perhaps you've seen people citing Claude or ChatGPT as a source during an online debate.
Last weekend, I saw first hand how blind trust in AI can lead people astray.
As I traveled to and from Washington DC last week, my connecting flights at the Dallas-Fort Worth airport were canceled or delayed twice due to severe weather- once on the way there and then again on my trip back to New Mexico.
Halfway through the flight from Washington DC to DFW, the pilot came over the intercom and announced that our flight would be landing in Memphis due to an unexpected, fast-moving thunderstorm in North Texas. When we touched down, everyone pulled out their phones and scrambled to find information about the storm and their connecting flights.
The gentleman in the seat next to me spoke into his phone and asked if there was a ground stop at DFW airport. Google's Gemini AI replied, "According to the latest information there is no active ground stop at DFW airport right now."
Meanwhile, I typed the search term "DFW ground stop" into my phone's Google app and scrolled past the AI-generated results. The first result was a Federal Aviation Administration website that said all flights into and out of DFW were halted.
Seconds later a flight attendant's voice came over the intercom and announced the DFW ground stop. I heard my neighbor mutter, "well that's not the information I'm getting." He seemed to trust Google Gemini more than the human flight attendant.
I leaned over and showed him the FAA website I found. Then I explained that as a journalist in a small newsroom, I'm used to finding ways to verify and double-check the information I have. I personally try to avoid asking AI for direct information, especially if in a fast-changing situation. Instead, I'll ask the AI where I can find certain bits of information. That way, I can verify things myself. Despite what the promise of AI might be, the technology still has a long way to go before it’s a trustworthy source of information.
