Researchers reveal a sharp rise in AI-generated spam, as cybercriminals use artificial intelligence to automate and refine email attacks.
A New Kind of Inbox Invader
Check your spam folder—it’s not just full of junk, it’s full of AI-generated junk. A new study reveals that over half of all spam emails clogging inboxes are no longer penned by humans, but by artificial intelligence tools.
Researchers from Columbia and the University of Chicago, in collaboration with cybersecurity firm Barracuda Networks, analyzed malicious and unsolicited emails from February 2022 to April 2025. Their findings show that AI-generated spam now accounts for 51% of all spam emails, a dramatic rise since the public launch of ChatGPT in November 2022.
Spam Gets Smarter, Not Just Louder
The study shows a steady growth in AI use across email threats:
- 51% of spam emails in April 2025 were AI-generated
- 14% of business email compromise (BEC) attacks used AI—though still mostly human-written
- AI-generated messages tend to be more formal, grammatically accurate, and cleverly phrased
- Cybercriminals are using AI to experiment with word choices to bypass security filters
According to the researchers, AI is primarily being used to enhance the content of spam—not radically change the type of attacks being carried out.
“By April 2025, the majority of spam emails were not written by humans,” said Prof. Asaf Cidon of Columbia University. “While targeted attacks like BEC still require a human touch, AI’s role is steadily growing.”
A Challenge to Detect and Defend
To determine AI’s role in email creation, researchers compared messages from before and after ChatGPT’s launch, training systems to detect AI-written content. Still, spotting AI involvement remains a challenge. “We can only see the output, not how it was created,” added Cidon.
India Not Immune to the Threat
Parag Khurana, Country Manager, India at Barracuda Networks, stressed the rising threat: “Cybercriminals are using AI to scale attacks. It’s crucial for Indian organisations to adopt platform-based defenses that offer deep threat visibility and rapid response.”
Barracuda recommends combining AI-powered email protection with cybersecurity training to guard against increasingly sophisticated spam and phishing attempts.
As AI continues to evolve, so too must our defenses because the spam of tomorrow might just be smarter than you think.