The Internal Revenue Service is warning taxpayers that artificial intelligence has given scammers a dangerous new edge, and with Tax Day approaching, the agency says the threat is more convincing than ever. AI-generated robocalls now mimic real IRS language so closely that even cautious filers could be fooled, ABC News reported.
The IRS flagged AI-enabled robocalls as part of its annual "Dirty Dozen" list of tax scams for the 2026 filing season. The agency published the list in a release last month, noting that "phone scams continue to evolve, including calls that use computer-generated tactics and spoofed caller ID to appear legitimate."
That is bureaucratic understatement. The real-world version is far worse.
Nomorobo, a company that specializes in identifying and blocking spam calls, intercepted an AI-generated phone call that was played on "Good Morning America" on Wednesday. The call used authentic IRS terminology, including the phrase "currently not collectible," a real agency status, and pressured the listener by claiming that "relief slots are limited, and they're filling up fast."
IRS spokesperson Yviand Hernandez told "GMA" that this kind of mimicry is now standard practice among scammers:
"Unfortunately, tax scammers are using AI to sound more realistic, use actual information, deductions, credits, or even programs that the IRS offers to taxpayers for either tax relief or even for collection."
The call sounded polished. It sounded official. And it was entirely fake.
For older Americans and retirees who may already be navigating IRS audit red flags or worrying about filing correctly, a call like that could easily prompt a panicked response, which is exactly what the scammers want.
The agency's own release laid out clear ground rules that every taxpayer should memorize. The IRS will generally contact people "by mail first." It does not leave "urgent, threatening prerecorded messages." It does not "call to demand immediate payment, or threaten arrest."
Any phone call that does those things is not the IRS. Period.
That distinction matters more now than it did five years ago. Before AI voice generation became cheap and accessible, most robocall scams were clumsy, stilted recordings, obvious foreign accents, garbled scripts. Today's AI-generated calls use natural-sounding speech, real tax terminology, and spoofed caller IDs that can display what looks like a legitimate government number.
The IRS "Dirty Dozen" list for 2026 also flagged fake charities and misleading tax advice on social media as threats to watch for during filing season.
The IRS warning did not arrive in a vacuum. The Federal Trade Commission told the Joint Economic Committee in prepared testimony last month that it had seen a "sharp increase" in the number of scam victims losing more than $100,000. Overall fraudulent losses climbed 430 percent over the last six years, the FTC said.
That number deserves a second look. A 430-percent increase in fraud losses is not a gradual uptick. It reflects a fundamental shift in the tools available to criminals, and a gap in the government's ability to keep pace.
The timing compounds the risk. Tax season is already stressful for millions of filers, and the 2026 season has brought its own headaches. An IRS policy shift delayed more than 830,000 refunds earlier this year, leaving many taxpayers frustrated and uncertain about the status of their returns. That uncertainty is exactly the kind of anxiety scammers exploit.
The IRS warning was not limited to incoming scam calls. The agency also cautioned taxpayers against using AI tools to prepare their own returns.
"Taxpayers should not rely on AI-generated responses to complex tax questions, and they should verify any calculations or information provided by artificial intelligence," the IRS release stated.
That is a significant admission. Millions of Americans now use AI chatbots for everything from drafting emails to answering medical questions. The temptation to ask an AI tool to help with a tax return, or to trust its output without checking, is real and growing.
The IRS is effectively telling taxpayers that the same technology criminals use to sound like the government can also produce plausible-sounding but wrong tax advice. For filers trying to navigate IRS operations and processing timelines during the 2026 season, that warning adds another layer of caution.
The practical takeaways from the IRS guidance are straightforward. If someone calls claiming to be from the IRS and demands immediate payment or threatens arrest, hang up. The real IRS does not operate that way. If a caller uses urgent language about limited "relief slots" or pressures you to act before a deadline, that is a scam, no matter how professional the voice sounds.
Do not trust caller ID alone. Spoofed numbers can display any label a scammer chooses. And do not hand over personal or financial information on any unsolicited call, email, or text that claims to be from the IRS.
If you receive a suspicious communication, the IRS has long encouraged taxpayers to report it. The agency's own channels, not a number provided by the caller, are the only safe way to verify whether a contact is legitimate.
There is a pattern here that should trouble anyone who believes government exists to protect citizens, not just to collect from them. AI has handed criminals a force multiplier, and the federal response so far amounts to press releases and annual lists. The FTC's own testimony shows fraud losses exploding. The IRS warns taxpayers to be careful. But where is the enforcement?
The "Dirty Dozen" list is useful as far as it goes. But a list is not a strategy. When scam losses jump 430 percent in six years and the best the government can offer is a reminder to check your mail, taxpayers are right to wonder whether anyone in Washington is keeping up with the people trying to rob them.
Criminals are using AI to steal from Americans. The least the government can do is match that urgency, not just with warnings, but with results.