When families weigh their options for paying for college, the question “What’s the best student loan?” seems straightforward. But as we learned after asking five leading AI models (ChatGPT, Grok, Perplexity, Copilot, and Google Gemini) the answers can be incomplete, sometimes unclear, and occasionally misleading.
Each tool offered a broad framework: start with federal loans, then consider private loans if needed. Yet critical details were often missing.
In some cases, explanations failed to note that Direct Subsidized Loans require demonstrated financial need. None clearly spelled out federal borrowing limits, an omission that could leave families unprepared if federal aid runs short. Or even specific scenarios or borrowing order (such as borrow the student's Direct Loan first, then supplement with Private or Parent PLUS Loans).
Last year, we discovered that Google's AI was getting financial answers incorrect 43% of the time.
When it comes to student loan recommendations, the results were better, but there were still key misses in an areas where policy change, interest rates vary, and borrower protections shift frequently.