AI shall not pass

TL;DR: For everyday people, the most compelling AI use case might be clawing back control over our devices and knocking out tedious errands. But several big companies are now shutting AI tools out of their products and services. It could be the start of a game of whack-a-mole: Consumers find a way to make AI work for them, and companies find a reason to nix it.

What happened: People keep finding novel uses for AI to remove some of life’s annoyances. The latest: vibe coding software to strip all ads and other bloatware from your smart TV. (Just look at this X user showing off a cleaner, less laggy TV screen.)

AI companies want you to make the most of their products (agentic or not), and users have taken them up on it—whether it’s by (allegedly) using ChatGPT to get out of a roughly $400 parking charge or deploying AI-coded software to expand what their existing music gear can do.

The burgeoning firewall: Unfortunately, it’s becoming harder to get AI to do what you want. Users recently noticed that Delta is now blocking third-party AI agents from booking flights, joining a growing list of services that have shut out AI bots (or at least appear to), including Amazon, Resy, and Amtrak. (Others, like Yelp, now make AI developers pay licensing fees for agent access.)

So far, companies have mostly pointed to security and customer experience as reasons for turning away AI agents. Amazon cited a “significantly degraded shopping and customer service experience” when it went after Perplexity’s shopping agent in a lawsuit late last year. (Perplexity argues there’s another reason the company takes issue with its agent: Fewer humans browsing Amazon also means fewer people seeing its ads.)

The security snag: And, to be fair, the AI blocks aren’t always intentional—sometimes agents just trip anti-fraud measures or get cut off for hammering a site with too many requests. Human customer service reps (you know, the few that still exist) report that they’re fielding a rising number of AI callers, which pose a problem because they can’t easily verify that a human authorized the bot, per Business Insider.

Bottom line: AI developers are going all out to convince people their AI tools are worth using—but an agent that struggles to book a flight or table may have people wondering what the tech is actually good for. —WK