Last week, OpenAI announced that ChatGPT is rolling out a new function for Pro users allowing them to link their bank accounts to the chatbot to get even more personalized financial advice.
There are already 200 million users asking ChatGPT money-related questions every month, according to an Endgadget article on the announcement. ChatGPT is teaming up with fintech firm Plaid, which already has a relationship with thousands of financial institutions. If you have a Venmo account, for example, there’s a good chance you used Plaid to connect it to your checking account.
Look, there’s something undeniably appealing about the ChatGPT offering. Managing your money can be complicated, and it sounds pretty cool to have a dashboard with all my financial information in one place and a computer bot on hand to answer all my questions, no matter how dumb they might be. The bots do tend to be overly solicitous, so you know it won’t be judgy like a real human might.
Still, my initial thought when I heard about this new function is that it’s a very bad idea. But then, as I read the marketing materials, I found my opinion shifting some. In theory, I’m all for the democratization of finance. I want it to be easier for people to get access to financial information and feel supported as they make difficult money decisions. The financial services industry has long relied on overly complicated jargon to keep people in the dark. Shouldn’t I be excited that the tech industry is finally making this information more digestible?
And yet all of it gives me the ick.
I worked on this essay all yesterday morning, and in the afternoon I traveled into Manhattan to attend Stefanie O’Connell’s book party. She’d invited a who’s who of the Lady Money Media Mafia™, and unsurprisingly, our conversations quickly turned to AI and this new ChatGPT function. Is it a good thing or a bad thing?
Truly, there’s nothing better than talking to smart women to help me figure out why I feel the way I feel on any given topic. As we talked, I warned them that our conversation would be informing today’s newsletter. And as I rode the subway home after the event, mulling over our conversation, this essay finally began to gel in my head.
I like technology. (Maybe I even love it?) I like to try new things and experiment. While I wouldn’t describe myself as an early adopter. I’m definitely not a luddite either.
And yet, I really don’t like generative AI. It’s been hard for me to explain why beyond making a few jokes here and there that I don’t want to train the robots for free. (I don’t.) But it’s so much more than that.
For starters, I don’t think the men running Anthropic and OpenAI have any respect for experts. They don’t see the value in spending years developing a craft and honing your skills. They have absolutely no remorse for stealing the knowledge it took humans generations to acquire and feeding it into their machines with little to no context.
Despite touting all that stolen knowledge, the bots are often wrong. The few times I’ve asked ChatGPT some basic personal finance questions, it made really stupid mistakes. Last year, it recommended Mint as a budgeting app. (Mint was shut down in spring 2024.) It frequently serves up outdated information on IRS rules and contribution limits. And once I tried to get it to create a chart for me on compound interest, and I realized it was compounding incorrectly. It can’t even do simple math right!
These might seem like small mishaps, but I sort of see it like Van Halen’s brown M&M rider amendment. You want whoever provides your financial advice to pay attention to even the smallest details. When putting together a financial plan, tweaking an investment forecast by even half a percent can have a huge impact on someone’s future. It could mean the difference between being able to retire in five years rather than 10.
But if you’re not an expert, and you’re turning to ChatGPT or Claude for advice, it can be increasingly difficult to check the veracity of the bots output. Have you tried to do a Google search lately? It’s complete garbage these days, and it’s only going to get worse as the company continues to prioritize Gemini, its own LLM product. On Tuesday, Google’s CEO announced it was overhauling the search bar function, making the dimensions bigger in an effort to encourage users to interact with it the same way they use Claude or ChatGPT.
It really scares me that it’s becoming so difficult to fact-check anything. And while I don’t want to sound like a total elitist, I doubt the average consumer is really taking the time to consider where ChatGPT or Claude is sourcing its information from and what its motives might be in offering you certain advice. These bots are not fidicuciaries—they are not required by law to put your best financial interest first.
Where I think many of us once saw the internet as a new frontier where a Google search could take you down a dozen different rabbit holes of discovery, this new technology is creating scary silos. Algorithms now dominate our lives, and Big Tech uses the data they’ve collected from us over the past 15 years to separate us into bubbles that feel impossible to escape.
Yes, the tech bros running these companies argue that AI is making our lives easier. If the algorithms know you better than you know yourself, you don’t have to make decisions. It will tell you what to eat, what to wear, how to raise your children, and what to do with your money. And you can use that extra time to chat with the bot who will tell you how pretty and smart you are or scroll an endless feed of 30 second videos that were chosen just for you.
I know I sound cynical as fuck. But I don’t like it here. I don’t like how polarizing the question of “to AI or not to AI” has become. Just because you’re the loudest voice in the room and you speak with great authority, it doesn’t mean you’re right. I believe that technology should work for humans, not the other way around. And I refuse to be bullied and belittled into believing otherwise.
So what’s the conclusion here? ChatGPT has a new personal finance function. Some people might find it helpful. As a financial expert, I would recommend you speak with a licensed professional before you make any big money moves. As a person who clearly has big feelings on this topic, I’d say there are better places to get advice.
-Lindsey
AI and personal finance in the news
- I’m a big fan of Andre Nader, who writes FAANG Fire, a personal finance newsletter geared toward tech workers. I reached out to him about this essay, and while I couldn’t fit his response into my final draft, I want to give him a shoutout. He’s a big AI fan, but I like him anyway! :)
- While there’s lots of noise about the need for more AI regulation, Mike Pierce explains in the IN DEBT newsletter that there’s (thankfully) already a lot of existing legislation which protects consumers from bad financial products. And while we might not get the help we need on the federal level, there are some very powerful state-level agencies that can serve as consumer watchdogs. Let’s hope they do!
- You’ll want to think twice before you ask AI to pick stocks for you: “Across a series of new trading contests between the world’s leading AI models, the verdict so far is unflattering. Most of the systems lose money. They trade too much. They make wildly different decisions when given identical instructions,” Justina Lee writes in a Bloomberg article (which I discovered via Matt Levine’s newsletter).
- “When people use our A.I.-powered features in search, they use search more,” Google CEO Sundar Pichai told The New York Times this week. This is funny to me because while Google touts it as a good thing, it seems like it could mean that people are not getting the responses they’re looking for with their first query. That’s certainly been my experience lately.
- One big thing I didn’t even have the word count to get into: What is OpenAI doing with my financial data beyond training ChatGPT? As Brian X. Chen wrote in The New York Times, we should pause to consider how much we’re sharing with the bots. “Unlike a traditional web search tool, chatbots invite people to type complete thoughts and follow-up questions, revealing their intentions much more explicitly.”
- Speaking of the bots getting things wrong, there was lots of buzz this week when the author of a new book about truth in the age of AI included quotes from experts that ChatGPT and Claude simply made up.
What else we read (and watched and listen to) this week
- I sobbed in the bathroom this morning listening to an edition of NPR’s StoryCorps featuring retired Army Major Duane Wittman speaking about the redbud tree planted in the memory of his son, Army Sgt. Aaron Xavier Wittman, who died fighting in Afghanistan in 2013. This Memorial Day weekend, I’m thinking of all the families who’ve lost children to war. -Lindsey
- I’ve been loving Kevin Morby’s new album, and I can’t wait to see him in concert in June! - Alicia
- I’ve seen quite a few people sharing a viral post written by a physician assistant recommending 26 different lab tests to ask your doctor for. As the daughter of a doctor and a pharmacist, I’m always a little skeptical of this kind of pop medical advice, and I really appreciated Istiaq Mian’s Substack Note offering a doctor’s take on the list. -Lindsey
- I’ve only read 13 of The Guardian’s 100 Best Novels of All Time. 🙈 What about you? - Alicia (I’ve read 25! And #1 is one of my all-time favorites! -Lindsey)
On our radar
- Alicia is looking to speak with a few DINKs (dual income, no kids) about weekend spending for Yahoo. If you’re interested in participating, reach out at alicia@thepurse.co.
Stat of the week
Data science employment is projected to grow 34% between 2024 and 2034.
-Bureau of Labor Statistics. Good news for anyone inspired by our recent high-earning data scientist Home Ec.
Comment of the week
“I’m 42, so an ‘elder millennial,’ and I can tell you it’s not just 30-somethings who feel this way. I’m solidly mid-career with around 20 to 25 years to go, and to think about the scale of change in that time to white collar work is a little head spinning. Then you add the creeping threat of ageism and ‘being too expensive’ and it gets pretty dark. My kids are 10 and 7 and it feels nuts to try and predict what employment will be like for them. And yet, big technological changes have happened before and something has emerged on the other side that has kept the workforce going. The impacts are not always evenly distributed (and I would argue we have way too much inequity baked in) but life does go on.”
- Lucy D on Alicia’s essay, “Where does the millennial career go from here?”
What else we published on The Purse this week
See how a mom is feeding four boys ages 4 to 14.
Alicia Adamczyk
Join the convo!
Alicia Adamczyk
What’s everyone going to do for the next 30 years?
Alicia Adamczyk
Don’t miss this great report from Kathleen Davis on the cost of summer camp.
Kathleen Davis
Best money we spent this week
- I spent a little time in Brooklyn on Sunday and enjoyed a meal outside at one of my favorite restaurants from back when I lived in Park Slope. So much has changed in the neighborhood, so I’m glad it’s still there! I read for a bit and spent a little time people-watching in the shade. Perfect! ($40) - Alicia
- My mom picks up Freddy on Tuesdays, and I’ve been using it as an excuse to grab a coffee with her first so we can have a few minutes to catch up. ($10 for two iced coffees.) -Lindsey
