A commercial phishing-as-a-service kit known as TheMailer is giving attackers a way to hand much of the campaign work to artificial intelligence. According to Dark Reading, the kit lets operators supply their own AI models, which are then used to personalize phishing lures, help messages evade spam filtering, and scale operations to tens of thousands of targets at once.

How the Kit Uses AI

The core shift is in who does the writing and tuning. Instead of relying on a template that a defender’s mail filter can learn to recognize, the attacker-supplied model generates message content tailored to each recipient. That variation undermines signature-based and pattern-based detection, which depends on repeated, recognizable text across a campaign.

  • Personalization: lures can be tailored to individual targets rather than reused verbatim.
  • Filter evasion: generated content is designed to avoid the patterns spam filters typically flag.
  • Scale: the kit is built to run campaigns against tens of thousands of recipients.

Why It Matters

Commercial kits lower the skill barrier for less sophisticated actors. When AI handles writing and evasion, a phishing operation that once required language skills and manual iteration can be run more cheaply and at greater volume. For defenders, this points toward a need to rely less on static content signatures and more on behavioral signals, sender authentication, link and domain reputation, and user reporting workflows.

What Security Teams Should Do

Organizations should review their mail security controls with the assumption that message text alone will be a weaker indicator going forward. Reinforcing multi-factor authentication limits the payoff of credential theft, while phishing-resistant authentication methods further reduce exposure. Security awareness training should also be updated to reflect that well-written, individually tailored lures are now cheap to produce.