- Kaspersky warns that AI is accustomed to create convincing emails, generate defects and more
- The number of clicks on phishing links grows fourth to quarter
- Users warned that they stay alert to be safe from attacks
The generative artificial intelligence (Genai) is making more intelligent, more difficult phishing attacks and has warned a more widespread new investigation of Kaspersky.
Their findings claim in the second quarter of 2025, their products detected and blocked more than 142 million clicks on Phishing links, which represents an increase of 3.3% in the first quarter.
While it may mean that there were more phishing attacks during the summer, it can also mean that the volume remained the same, but the real attacks became more convincing.
Generating deep defects
“The AI has raised phishing to a highly personalized threat. Large language models allow attackers to create emails, convincing messages and websites that mimic legitimate sources, eliminating grammatical errors that once exposed scams,” the researchers said.
“IA bots driven in social networks and messages are passed through real users, involving victims in prolonged conversations to generate confidence. These bots often feed romantic or investment scams, attracting victims to false opportunities with audio messages generated by deep defake videos.”
The researchers also said that Genai is not used only to eliminate spelling and grammar errors, possibly the largest red flags in Phishing’s emails.
It is also used to create audio impersonations and realistic videos of trusted individuals, including colleagues, celebrities and bank officials.
These deep are used to promote false gifts or steal confidential information, such as multifactor authentication codes (MFA), passwords and the like.
“The tools with AI analyze public data from social networks or corporate websites to launch specific attacks, such as emails with human resources issues or false calls that refer to personal data,” Kaspersky added.
How to stay safe
As Phishing’s risk grows, staying online is still the best way to be sure.
Users should always be skeptical of unre requested incoming messages, especially those that demand urgent actions or threaten a disaster.
These are, and will continue to be, the largest red flag in Phishing attacks.
