A familiar voice on the phone is becoming a less reliable basis for trust as fraudsters use artificial intelligence to impersonate relatives, a senior cybersecurity researcher has warned in an interview about Bangladesh.
Speaking to The Daily Star, Sergey Lozhkin, who heads Kaspersky’s Research Centre for APAC and META, described scams that combine synthetic voices or images with an urgent request for money.
Urgency is part of the deception
The scenario can sound personal: a caller claims a family member has suffered an accident and needs immediate help. Lozhkin said such tactics exploit fear, including among people who already know impersonation scams exist.
What the threat figures actually measure
Kaspersky reported detecting more than 2.18 million internet-borne threats in Bangladesh between April and June 2026. It said 23.4 per cent of its users in the country encountered web threats.
Those figures describe the company’s participating user base and detections. They are not a count of all Bangladeshis affected, nor a measurement of money lost to AI fraud.
Lozhkin also argued for stronger institutions, public awareness and technical training. For young researchers, his emphasis was on programming, reverse engineering and understanding how systems work.
For readers, the practical distinction is between recognising a voice and independently confirming a request. A separate call to a known number creates a chance to check the story before urgency turns into a payment.
