AI chatbots push premium credit cards 5.7x more than cheap ones — and pull from the same few affiliate sites

Becky Robertson

Fri, July 3, 2026 at 2:30 AM PDT 4 min read

Despite warnings against relying on AI for financial (and other) advice, a growing number of Americans are comfortably turning to chatbots to tell them what to do with their assets.

Yet another analysis has shown the bias large language models tend to have, serving as another reminder to always apply a reasonable level of skepticism to their guidance, which can and often does come from a narrow group of sources.

Must Read

• Jeff Bezos backs a platform that lets anyone invest in rental homes for as little as $100 — 6 ways to build wealth like a landlord without actually being one

• Robert Kiyosaki says this 1 asset will surge 400% in a year and begs investors not to miss this 'explosion'

• Millionaires under 43 hold only 25% of their wealth in stocks. Here's where their money is actually going

The study (1), conducted by New York-based AI communications firm 5W and titled the Credit Cards AI Visibility Index, tested out prompts that consumers might use when researching which credit card is best suited to their money goals. And, in response to thousands of queries, it found that the most-utilized AI platforms drew from a very high concentration of websites — and often recommended some of the most expensive options as a result.

Blatant provider advertising isn't vastly skewing recommendations — but other forms of marketing might be

As the company noted, credit cards are the most aggressively promoted financial products in the country, soaking up $20 billion of bank marketing budgets each year. However, the investigation found that "when American consumers ask AI engines which card to apply for, the answer is not supplied by the issuers" — so, answers were not heavily influenced by direct advertising — the same small pool of resources was used again and again to inform the suggestions the chatbots provided.

The bulk of citations in AI responses (60-65%) came from a small collection of sites, including The Points Guy, NerdWallet, Bankrate, CardRatings, WalletHub and Forbes Advisor.

While ChatGPT and others treated these sites as "default sources of truth" on the subject, 5W says, it's not because they are inherently the most reliable, but because of factors like the volume of posts they have about credit cards, how often they publish new posts and the style their posts take (such as comparing different credit cards side by side).

The study also points out that many of these sites earn commission when readers sign up for a credit card, and they might choose to highlight more expensive cards as a result. Most sites that engage in affiliate marketing, including Moneywise, have policies that state commissions do not affect cards reviews and rankings.