We’re at a surreal stage of generative AI adoption, as some of the growing pains of this still relatively new technology are revealed in funny and bizarre ways.
Google released “AI Overviews” at scale in the US recently, giving everyone an opportunity to kick the tires and ask Google questions answered by AI, which then appeared at the top of the search results.
Some of the answers were hilarious and strange, like its suggestion to “eat one small rock a day”, or its assertion that “a dog was a fourth-round NHL draft pick and played 63 games for the Calgary Flames,” or its advice that “glue can help cheese stick to pizza.”
My favorite recent AI growing pains story comes from the Chevrolet Watsonville car dealership, which introduced an AI chatbot to ask web site visitors, “how can I help you in your vehicle search?”
Website visitors soon had the Watsonville Chevrolet chatbot saying all sorts of crazy things, from praising Tesla to writing Python code. Chris Bakke even convinced the chatbot to agree to sell a 2024 Chevy Tahoe for $1 and had the chatbot say “and that’s a legally binding offer – no takesies backsies.”
These are all part of the awkward adolescence of AI and digital transformation, which I’m speaking about this week at the Gartner Marketing Symposium/Xpo in Denver.
It’s a particularly fitting event, given Gartner’s famous Hype Cycle for emerging technologies. Last August, Gartner pegged GenAI at the “Peak of Inflated Expectations” in the Hype Cycle. What comes next is the “Trough of Disillusionment,” which many businesses are experiencing as they figure out how exactly to navigate the promise of GenAI.
As Rumman Chowdhury at Humane Intelligence said recently:
“No one wants to build a product on a model that makes things up.”
It will take time and experimentation for brands to figure out how best to leverage AI beyond the hype.
Here are a few related cartoons I’ve drawn over the years:
