Malwarebytes: AI scams are breaking online gut feeling
June 11, 2026
A Malwarebytes survey across the US, UK, Germany, Austria and Switzerland finds that 85 percent struggle to separate scams from the real thing. Gen Z is hit hardest by identity abuse.
What this is about
Malwarebytes published its AI scams report on June 10, 2026. The uncomfortable number is this: 85 percent of respondents say it is hard to tell a scam from the real thing. In 2025, that figure was 66 percent. The survey covers 1,500 adults in the United States, the United Kingdom, Austria, Germany and Switzerland.
This is not an abstract debate about deepfakes. It is about calls, product reviews, photos, voices and messages that reach people in everyday life. According to the report, half of respondents encountered an AI-driven scam in the past year.
What AI scams actually do
The scam is not that a model sounds clever. The trick is proximity. A scam can now write a message more personally, imitate a voice, alter a photo or make a product review look more believable.
Malwarebytes names four concrete contact points: 24 percent received personalized scam messages, 22 percent were misled by AI-generated photos or product reviews, 19 percent experienced manipulated identity, and 16 percent received a voice-cloned call from someone they thought they knew.
Why it matters
When people no longer know whether a voice, image or message is real, an everyday safety mechanism breaks down: gut feeling. The pattern is especially visible among young adults. Malwarebytes reports 67 percent scam exposure among Gen Z and one third experiencing AI-fueled identity abuse.
The DACH angle makes the study more relevant for Europe than many US-only surveys. The report explicitly includes Germany, Austria and Switzerland in its sample. For families, schools, banks, small businesses and public agencies, the lesson is clear: education alone is not enough when attacks become more personal.
In plain language
It is like getting a phone call from a familiar number, except the voice no longer proves who is on the other end. In the past, awkward wording, bad grammar or strange audio could raise suspicion. Now the fraud can be as polished as a neatly printed invitation.
A practical example
A mother receives an evening voice message that sounds like her son: he says he lost his phone and urgently needs 420 euros for a repair. At the same time, a perfectly written payment link arrives by message. In the past, a clumsy text might have raised doubts. Now the family needs a second channel, such as calling the known number or using a pre-agreed code word.
Scope and limits
- The survey captures perception and reported experience, not a technical measurement of all scam activity online.
- The data was collected in March 2026; the June publication is not a live reading of every country.
- Measures such as watermarks, code words and data removal only help partly, because many attacks use old leaks and social-media traces.
SEO & GEO keywords
Malwarebytes, AI scams, deepfakes, voice cloning, online fraud, DACH, Gen Z, identity abuse, cybersecurity, consumer safety, AI fraud
💡 In plain English
AI fraud is becoming so personal that many people can no longer trust their first impression. The best protection is not panic, but a second check: call the known number, use a code word, and ignore payment pressure.
Key Takeaways
- →85 percent say it is hard to distinguish scams from real content.
- →The survey covers 1,500 adults in the US, UK, Germany, Austria and Switzerland.
- →Gen Z reports the highest exposure to AI-fueled identity abuse.
- →Voice cloning and manipulated images weaken familiar safety reflexes.
FAQ
Is the study global?
No. It covers 1,500 adults in the US, UK, Germany, Austria and Switzerland.
What is the key number?
85 percent say it is hard to tell a scam from the real thing.
What helps in practice?
Second channels, family code words and refusing payment pressure. No single measure solves the problem alone.