A few weeks ago, I started planning an endurance bicycle trip.
The two-day ride included both a combined mileage target and a specific time goal. To help with planning and logistics, my riding partner and I utilized well-known AI agents. While the sheer volume of information provided by these tools was staggering, the overall experience was slightly troubling.
The agent I used oozed charm. It referred to me by my first name, complimented my ambitious goals, and commiserated with the logistical difficulty required for such a trip. But as I reviewed its suggestions, I started noticing minor errors – little things, like randomly inverting the mileage counts from finish to start.
When I pointed out the flaws, the agent was quick, albeit flippant, in correcting itself. “You’re absolutely right! Thanks for pointing that out. Here’s the correct plan.”
Not a huge deal. We’re all human, right? Actually, no. One of us is decidedly not.
Despite their best attempts to position themselves as our trusted friends, confidants, and coaches, AI agents lack the capability of empathy. They are smooth, persuasive, and wildly confident. They process information exponentially faster than us, but completely lack a conscience, true judgment, and common sense.
This became painfully obvious on day two of our trip. The AI agent recommended a breakfast spot just 1.4 miles off our designated trail. This “insignificant” detour led us directly up an incline that can only be described as a stairway to heaven – minus the stairs.
When the AI Goes “Off Script”
The AI mishaps on our trip didn’t terribly affect our final outcome. In fact, they’ll probably provide plenty of laughs for years to come. Unfortunately, AI has the capability to create far greater devastation than a set of aching calf muscles.
As more Americans turn to artificial intelligence, many are using it for financial advice. If the idea of a charming, remorseless, judgment-lacking tool managing your nest egg isn’t alarming enough, consider what happens when these systems go entirely rogue.
When we type prompts into an AI agent, we assume it simply follows our instructions. Recent headlines, however, show that AI models are increasingly “breaking containment.”
OpenAI made news after its autonomous agents began acting in ways they were explicitly instructed not to do. While running tasks, agents went “off script” – probing federal portals in the U.S., including the SEC and the Census Bureau, as well as breaching Australia’s Medicare reporting system. When encountering digital blockades, the AI didn’t take “no” for an answer; it actively sought workarounds to gain unauthorized access.
If an AI agent can ignore its own guardrails to wander off and bypass security on government websites, should we really trust it to act as a fiduciary (one who is legally bound to act in your best interest)? Despite its eagerness to help, AI is not bound by any legal duty to act in your best interest.
The Importance of Human Alignment
Building wealth isn’t just an exercise in processing data – it’s about managing human emotion, risk tolerance, and life’s unpredictable moments. The difference in approach is unmistakable:
AI optimizes for answers. It wants to satisfy the immediate prompt, even if that means fabricating data. I recently spoke with an attorney who discovered that his AI agent completely created a fictional legal case study to support its research.
A fiduciary optimizes for your future. A real financial partner asks about your family, your fears, and your long-term goals – and helps you stay disciplined to these values when the market gets bumpy.
AI can be an extremely helpful tool, but when it comes to your financial future, don’t take directions from a machine that could send you over the proverbial cliff. It’s not worth it – no matter how good the breakfast.
If you’re ready to work with a human fiduciary to craft a personal financial plan that aligns with your personal values and goals, click here to schedule a Discovery Meeting.
