Security researchers have reported a campaign in which attackers use Google Search ads together with legitimate Bing redirect URLs to steer users toward fake Claude installer pages. The pages are designed to deliver a ClickFix-style attack, a social engineering technique that tricks users into running harmful commands or malware under the guise of fixing a problem or completing an installation.

What happened

According to the reported findings, threat actors placed malicious Google ads that used Bing search-result redirect links as click URLs. Because the links point to a legitimate Microsoft-owned domain during part of the journey, the campaign may appear more trustworthy to users and can be harder to spot at a glance.

Users who followed the ads were reportedly redirected to fake pages impersonating Claude, the AI assistant. Those pages then attempted to deliver a ClickFix attack through a fraudulent installer flow. This is part of a broader pattern in which attackers use interest in popular AI tools to attract clicks and compromise devices.

Why it matters for European businesses

This incident is relevant beyond the specific brands involved. Many European businesses are actively testing or deploying AI tools, and employees may search for assistants, browser tools, desktop clients or integrations without going through approved procurement or IT channels. That creates an opportunity for attackers to abuse paid search, lookalike websites and installer workflows.

The campaign also shows how legitimate advertising and redirect infrastructure can be misused in a way that reduces suspicion. For businesses, this increases the risk that staff may trust a result simply because it appears as a sponsored listing or because part of the URL path involves a recognised platform.

For organisations with distributed teams, bring-your-own-device policies or limited endpoint controls, this kind of attack can lead to credential theft, malware infection or unauthorised remote access. The operational impact can include disrupted work, incident response costs and possible data protection consequences if business or customer information is exposed.

Who may be affected

The risk is especially relevant for:

• SMEs and founder-led businesses that rely on web search to discover and install new tools quickly.
• Marketing and content teams that frequently test AI assistants and productivity software.
• IT and security teams responsible for browser security, endpoint protection and software approval processes.
• E-commerce and digital businesses where compromised employee devices could expose admin accounts, ad platforms, analytics tools or customer data.
• Agencies and service providers whose staff manage multiple client environments and could create wider downstream risk if accounts are compromised.

What companies should consider

• Restrict software installation paths: Encourage staff to access AI tools through official vendor websites, known bookmarks or approved app catalogues rather than through ads or general search results.
• Review endpoint protections: Make sure endpoint security tools can detect suspicious scripts, command execution and malware delivered through fake installers.
• Train staff on ClickFix tactics: Employees should be cautious if a website asks them to copy commands, disable security controls or run unexpected installation steps.
• Use browser and DNS controls: Web filtering, safe browsing features and DNS security services can help block known malicious domains before a user downloads anything.
• Apply least-privilege access: Limiting local admin rights can reduce the damage if a user launches a malicious installer.
• Monitor AI-tool adoption: Keep an inventory of approved AI services and communicate clearly how employees should access them.
• Check account exposure: Where possible, strengthen access to email, cloud admin consoles, ad accounts and collaboration tools with phishing-resistant multi-factor authentication.

While the reported campaign centres on fake Claude installers, the underlying lesson is broader: interest in AI software is now a reliable lure for attackers. European businesses adopting AI should treat discovery, downloading and onboarding of new tools as a security process, not just a productivity decision.