Law enforcement and industry now say criminals are using artificial intelligence to supercharge phishing, voice clones, and online fraud at a scale that strains basic defenses.
Story Snapshot
- The Federal Bureau of Investigation (FBI) warns criminals are using artificial intelligence for advanced scams and deepfake voice and video cloning.
- Microsoft’s 2024 report says artificial intelligence is reshaping cyber threats, including more targeted phishing and impersonation.
- The FBI’s Internet Crime Complaint Center reports record losses tied to online fraud in 2024.
- Researchers find artificial intelligence lowers the skill and cost needed to run convincing cyber scams.
What Changed: Artificial Intelligence Supercharges Old Scams
Federal Bureau of Investigation officials in San Francisco warned in 2024 that criminals are using artificial intelligence tools to run smarter phishing and social engineering, plus voice and video cloning scams. Microsoft’s 2024 Digital Defense Report adds that artificial intelligence is reshaping the threat picture, from targeted email lures to deepfake content used to mislead staff and customers. Together, these signals point to a clear shift: old tricks now move faster and hit more people with fewer mistakes.
The core change is scale and speed. Artificial intelligence can write emails that sound natural, fix grammar, and mimic a company’s style. It can also fake a boss’s voice to rush a payment or pose as a customer to reset an account. These tasks used to take time and skill. Now they can be done in minutes by people with little training. That makes every inbox and help desk a bigger target, more often, with better bait.
The Numbers: Growing Losses and Expanding Access
The FBI’s Internet Crime Complaint Center reported online crime losses of $16.6 billion in 2024, a one-third jump from 2023, underscoring the rising cost to families and businesses. The Alan Turing Institute’s Centre for Emerging Technology and Security reports “considerable evidence” that artificial intelligence use in crime is accelerating, especially in phishing and financial fraud. These findings match what many teams see on the ground: criminals reach more victims because artificial intelligence makes fraud content cheaper and more believable.
Industry researchers have also tracked custom criminal chatbots designed to help with business email compromise, a scam that tricks employees into sending money or data to the wrong place. While vendor reports vary in depth, they describe a consistent pattern: artificial intelligence lowers barriers to entry, boosts volume, and increases personalization. That is enough to overwhelm busy workers and small security teams. Even minor gains in click rates or response rates can add up when attackers run thousands of attempts a day.
Why It Resonates: A System Strained by Automation
Small businesses, city agencies, and schools face the same bots as big firms but with fewer tools and staff. Many cannot train workers fast enough, patch systems on time, or verify voices on a tight deadline. When artificial intelligence removes friction for criminals, the burden shifts to the rest of us to slow down and double-check. That is hard when daily life pushes speed over care. People across the political spectrum see a common theme: systems that are supposed to protect the public often move too slowly to match automated threats.
Public trust also takes a hit. Deepfake voices and videos can shake faith in what we hear and see. That matters for business, local government, and daily life. A fake call from a “CEO,” a “grandchild,” or a “pastor” can cause panic and fast mistakes. The Federal Bureau of Investigation warning stresses this risk and urges careful verification before sending money or data. The message is simple but tough to follow under pressure: pause, confirm, and use a second channel.
What Defenders Can Do Right Now
Security teams can raise the floor with a few steps. First, train for the new tactics, including voice clones and urgent payment requests. Second, add simple checks like call-back rules, no exceptions. Third, filter and flag lookalike domains and newly registered websites. Fourth, use email authentication and outbound warnings for external senders. These steps do not stop every attack, but they cut the quick wins that artificial intelligence now enables at scale, buying time to respond.
Prompt injection is attractive because the attacker can make the victim’s own automation do the expensive work.
Instead of building infrastructure to steal, sort, and interpret data, an attacker may manipulate an AI-enabled workflow that already has access, context, and compute.… pic.twitter.com/KjN1rxJI5J
— Guardian Digital, Inc. (@gdlinux) August 14, 2026
Leaders should also match automation with automation. Many tools can spot unusual patterns, strange writing styles, or risky links in real time. Microsoft’s analysis highlights a two-sided race, where artificial intelligence helps both attackers and defenders. The goal is not perfect safety. It is to make each step of a scam harder, slower, and more costly. When that happens, many criminals move on to easier targets. That shift protects schools, clinics, and small firms that keep communities running.
Sources:
newscientist.com, microsoft.com, trendmicro.com, dhs.gov, securityboulevard.com
© truetrendnews.com 2026. All rights reserved.
