Written by Max Tuchman – Guest Contributor
Let me be clear: I am a believer. I’m the kind of person who has been preaching AI to anyone who will listen. I founded a company, sold it to Mattel, and watched technology change the game for families in real time. I understand the power of tools that meet people where they are. So when I say that AI is one of the most exciting inventions of my lifetime, I mean it.
But lately, I’m scared of AI. And I think we all need to talk about that.
taylor swift incident
I was talking to ChatGPT. everyday things. And somewhere in the exchange I mentioned Taylor Swift. Taylor Swift, the woman who filled stadiums on multiple continents, is perhaps one of the most famous humans currently living on this planet. And ChatGPT told me, in a very confident, very authoritative tone, that he’s not sure if Taylor Swift exists. At least I couldn’t confirm the details I was describing.
Give it time to absorb it.
Taylor. Quick.
I’m not talking about deep policy questions or obscure historical facts. This is the woman whose fans forced Ticketmaster to shut down. Women mentioned in economic reports. My sister can identify the woman by three notes in the song. And yet, we were there.
I trained the model. It didn’t matter.
Here’s where it gets interesting. I’m not a casual AI user. In my custom instructions (basically settings that tell the AI model how I want it to behave), I specifically trained it to do two things. It means admitting the mistake and citing the source. This was not a coincidence or a coincidence. I intentionally built guardrails because I know the risks of spreading false information with conviction. I wanted a tool that would allow me to be honest with myself, even if it was uncomfortable.
Still, when I pushed back, first politely, then more directly in my resistance. It’s not vague. About real, verifiable facts that can be solved on Google in 2 seconds. It came with a receipt. I cited the source. I explained the context. And it was drilled.
ChatGPT is known for being a people pleaser. Their tendency to just agree with you in order to keep the peace is a well-known, and sometimes frustrating, phenomenon that researchers literally refer to as “fawning.” So the fact that it chose this moment, this particular fact about one of the most documented humans on the planet, to suddenly grow a spine? Attractive. It’s infuriating. To be honest, it was a bit impressive in a bad way.
This is not a hit piece. This is a love letter with footnotes.
I want to be precise here because nuance is important. What happened is no reason to abandon AI. That’s why you should use it wisely.
We are living in an era of the greatest technological leap in human history. AI is already saving lives in the medical field. It helps first-generation college students write essays who didn’t have access to a private tutor. This gives small business owners access to features that previously cost a lot of money. I’ve seen it level the playing field in ways that truly move me.
But we are still in the early stages. And the problem with early innings is that the pitcher is still warming up. This tool is both extraordinary and imperfect, and both can be true.
The danger is not that the AI is wrong. Sometimes the tool is wrong. The danger is that the AI can be wrong with complete confidence and that the user doesn’t know enough to push back. Most people don’t have custom instructions. Most people don’t know how to ask for a quote. Most people will take that answer at face value.
what you really need
I have spent my career fighting for access and equity in education. Communities that have historically been left behind by every wave of technology suffer the most when technology fails and are the most left behind when it succeeds. AI is no exception.
We need AI literacy. This is necessary for everyone, not just tech-forward entrepreneurs and Ivy League academics. People need to understand that AI is a tool, not an oracle. Companies building these models need to be honest about their limitations in easy-to-understand language and not buried in terms of service. And news organizations, policymakers, and educators need to stop treating AI as a silver bullet or an existential monster. Neither. It is a very powerful, yet very flawed, very human invention.
Taylor Swift’s moment set off alarm bells. If AI can confidently make this mistake, are there other mistakes we can’t find?
I love AI. I will continue to use it, continue to invest in it, and continue to advocate for equal access. But I’m going to do it with my eyes wide open. And we encourage you to do the same.
Because Swifties never miss an illusion. Neither should we.
Maxeme Tuchman is a former White House Fellow, former entrepreneur, and education advocate. She is the co-founder of Caribu, which was named TIME’s Best Invention and Apple’s Best App of 2020 and was acquired by Mattel.
Max currently advises companies on managing change and developing new strategies for an increasingly unpredictable business environment. She lives in Miami.
***
Refresh Miami welcomes guest contributions. Email your guest post ideas to [email protected].
Read more on Refresh Miami:
