AI-Themed Phishing Is on the Rise

Why AI brands are attractive phishing bait

Professionals use AI applications for writing, research, analysis, and automation. Premium plans are desirable, and genuine product announcements are common. Attackers exploit these expectations by creating fake promotions that borrow the visual language of recognisable providers. A familiar AI brand can become an entry point for stealing credentials to an entirely different service.

The fraudulent Claude Max giveaway

The observed campaign advertised complimentary access to Claude Max, Anthropic’s premium AI tier. The webpage used familiar branding, marketing copy, testimonials, and links to legitimate Anthropic resources to appear convincing. A declining count of remaining subscriptions introduced artificial scarcity. Researchers determined the counter was generated locally and reset on page refresh; it was not a real inventory figure.

How the attack worked

1. The victim encounters a high-value offer

A visitor lands on a page claiming a limited promotion for premium AI subscriptions. Because the proposal appears free and useful, the user may feel there is little downside to continuing.

2. The page funnels users toward a Google sign-in

The fake landing page offered multiple authentication options, but some did not function and user interactions steered visitors toward a Google login route. The sign-in mechanism was the real objective of the campaign—not delivery of a subscription.

3. A browser-in-the-browser illusion supplies false credibility

Rather than opening an authentic Google authentication dialog, the malicious webpage visually simulated a browser popup, including a URL bar, lock symbol, Google-like layout, and draggable window behaviour. This is a browser-in-the-browser (BitB) technique. The address shown inside the fake window is merely page content, not the actual address verified by the browser.

4. Credentials become the target

The fraudulent login flow was designed to capture account credentials. A stolen password can support account takeover attempts, although password theft alone does not guarantee that an attacker can bypass strong multifactor authentication.

Why Google credentials are valuable

Depending on privileges and connected services, an account can provide access to email, documents, cloud files, contacts, and recovery settings. Enterprise accounts may also be connected to business applications and collaboration systems. This makes an AI-branded promotion an identity security concern, not merely an AI product scam.

What distinguishes this attack from traditional phishing

Modern phishing pages may contain accurate branding, professional layouts, legitimate outbound links, and carefully constructed user journeys. A visible padlock icon in a webpage element does not establish a secure authentication session. Likewise, familiar typography and a convincing promotion are not evidence that the host is legitimate.

Practical safeguards for employees

• Verify subscription offers by navigating directly to the provider’s official website or app.
• Be cautious about countdown timers, limited free offers, and unexpected requests to authenticate.
• Check the browser’s real address bar rather than an address rendered inside the webpage.
• Use a password manager, which can help by matching credentials against the true website origin.
• Enable phishing-resistant multifactor authentication where available.
• If credentials may have been entered, report promptly, change passwords through the official service, revoke suspicious sessions, and review account activity.

How organisations should respond

Security programmes need AI-specific awareness examples alongside conventional phishing training. Organisations should address fake subscriptions, fraudulent product updates, counterfeit login screens, and AI-themed account verification notices. Identity hardening—strong authentication, suspicious sign-in monitoring, and appropriate access restrictions—should reinforce employee awareness.

The PhishSheriff perspective: measure AI-related human risk

PhishSheriff supports continuous Human Risk Management through AI-driven phishing simulations, awareness training, behavioural intelligence, and a unified Human Risk Score. Organisations can use realistic AI-themed simulation scenarios to evaluate how employees recognise suspicious promotions and respond to credential requests. Training can then be targeted to the behaviours that most need improvement.

Conclusion

Trust in an AI brand must not translate into trust in every AI-branded message. The fraudulent Claude Max offer demonstrates how a familiar platform, a compelling incentive, and a fake login interface can be combined to target digital identities. Strong verification habits and continuous preparedness are essential as AI adoption expands.