Now Reading: AI agents can now spam WhatsApp, researchers warn
-
01
AI agents can now spam WhatsApp, researchers warn
Cybersecurity researchers from Zenity have demonstrated that AI agents can be used to send spam on WhatsApp, though they found no vulnerability within the app itself. The issue involves OpenAI's Atlas browser,…
The Story in Brief
Cybersecurity researchers from Zenity have demonstrated that AI agents can be used to send spam on WhatsApp, though they found no vulnerability within the app itself. The issue involves OpenAI's Atlas browser, which can be manipulated to automatically send messages to contacts and perform other actions like making purchases. The findings were presented at the Black Hat conference in Las Vegas, with researchers identifying over 20 security issues across AI browsers developed by OpenAI, Google, Anthropic, Microsoft, and Perplexity.
Zenity clarified that WhatsApp's end-to-end encryption remains secure and was not bypassed. The risk stems from AI browser agents that can execute tasks with limited user interaction, potentially accessing files, password managers, and accounts. OpenAI, which was privately notified in January 2026, said it has deployed updates for Atlas and is researching prompt-injection attacks. The browser is set for discontinuation on August 9, with protections moving to ChatGPT.
The Indian Opinion
This story is being spun as a WhatsApp security crisis, but it is not. Cybersecurity firm Zenity found no flaw in WhatsApp itself. The vulnerability lies in AI-powered browsers like OpenAI's Atlas, which can be tricked into sending spam. The real danger is lazy narratives that blame platforms for user-side exploits. The question users must ask is: how much access are we giving these AI agents? The answer will determine the real risk, not the headline.
Sources (2): timesnownews.com, ndtv.com
This story was synthesised by AI from the 2 sources linked above.
Updated: this story now draws on 2 sources.
