Security reporting indicates that ClickFix attacks are growing, with attackers increasingly using fake AI software installers, manipulated search results and trusted online platforms to convince users to execute malicious commands or files on their own devices. The reported trend is notable because it combines social engineering with interest in AI tools, making the attack path more believable for employees and harder for organisations to block with traditional controls alone.

What happened

According to reporting on research from Push Security, cybercriminals are scaling ClickFix-style attacks by presenting malware as legitimate fixes, downloads or setup steps. In these campaigns, victims may encounter fake AI tool installers, poisoned search results or malicious instructions delivered through environments that appear familiar and trustworthy.

The core tactic is not new: instead of relying only on a hidden exploit, attackers persuade the user to take an action that bypasses normal caution, such as copying a command, running a script, approving a download or installing software that appears useful. What appears to be a productivity or AI-related tool can in fact lead to malware execution, credential theft or broader compromise.

Why it matters for European businesses

This matters because many European companies are actively testing generative AI tools, browser extensions, desktop clients and workflow utilities. That creates an opportunity for attackers to exploit urgency, curiosity and limited internal governance around which tools staff are allowed to use.

For SMEs in particular, the risk is practical rather than theoretical. Employees may search for an AI assistant, meeting summariser, image generator or coding tool and click a promoted or manipulated result. If the user is then instructed to run a command or install a package manually, the attacker may bypass some preventive controls by making the action appear voluntary and work-related.

These attacks can lead to several business problems:

• Endpoint compromise: malware can gain access to laptops and workstations used for email, finance, CRM or internal systems.
• Credential theft: stolen browser sessions or passwords can expose Microsoft 365, Google Workspace, cloud services and e-commerce back offices.
• Data exposure: customer data, internal documents or commercially sensitive information may be accessed or exfiltrated.
• Operational disruption: infected devices can lead to downtime, incident response costs and wider remediation work.
• Compliance consequences: if personal data is affected, companies may face GDPR-related notification and assessment obligations depending on the incident.

The use of AI branding as a lure is especially relevant for companies that have not yet formalised procurement, approval and security review for AI tools. Interest in rapid adoption can create a gap between business demand and IT oversight.

Who may be affected

The reported trend is relevant across sectors, but some groups may be more exposed:

• SMEs and founder-led businesses with limited security teams and informal software approval processes.
• Marketing and content teams that frequently test new AI assistants, image tools and browser-based services.
• Developers and technical staff who may be asked to run commands, scripts or local packages as part of setup.
• E-commerce businesses where compromised admin accounts can affect storefronts, orders, customer data and payment-related operations.
• Remote and hybrid organisations that depend on cloud identity, browser sessions and SaaS access from unmanaged or lightly managed endpoints.

What companies should consider

Businesses do not need to stop AI adoption, but they should reduce the chance that staff install unapproved tools or follow unsafe setup instructions.

• Define an approved AI tool list: give employees a clear route for using sanctioned AI services and software.
• Limit local admin rights: reduce the ability of users to install software or run high-risk commands without review.
• Strengthen endpoint protection: ensure laptops and desktops are monitored for suspicious script execution, unusual downloads and persistence mechanisms.
• Train staff on the current tactic: update awareness guidance to cover fake AI installers, search poisoning and prompts to copy commands into terminals or system tools.
• Use safer software distribution: direct employees to official vendor pages, internal app catalogs or managed deployment tools rather than ad-driven search results.
• Review browser and identity protections: protect cloud accounts with phishing-resistant MFA where possible and monitor for unusual logins or session abuse.
• Prepare an incident process: make sure employees know how to report a suspicious install attempt quickly so devices can be isolated before wider damage occurs.

The broader lesson is that AI adoption is also a security governance issue. As new tools enter daily workflows, businesses should treat discovery, testing and deployment as controlled processes rather than leaving software choices entirely to end users.