Cyber Blog

Cybercriminals are increasingly pretending to be popular artificial intelligence (AI) tools to trick employees into downloading harmful software, giving away passwords, and exposing business accounts. These scams often imitate familiar services such as ChatGPT, Claude, Google Gemini, and Microsoft GitHub Copilot. According to cybersecurity research from ANY.RUN, 1,258 public security investigations were linked to fake AI campaigns between April and October 2026. The United States accounted for 16% of the targets where a country could be identified, highlighting the growing risk to American businesses.

These scams work by taking advantage of people’s trust in well-known AI products. Criminals create fake websites, advertisements, and download pages that look like legitimate services. An employee searching for an AI tool might click on a sponsored search result or visit a website that appears genuine, only to be directed toward a scam. The goal may be to steal login details, gain access to company accounts, or secretly install software that collects sensitive information from the victim’s computer.

One example involved a fake website pretending to offer an official download of Claude, an AI assistant. Visitors were shown a download button that appeared to install the expected application. However, clicking the button copied a harmful command onto the user’s computer and instructed them to run it using a Windows tool called PowerShell. Instead of installing the legitimate AI application, the command started a chain of actions that downloaded and launched malicious software. This allowed the attackers to carry out their attack while making the process appear like a normal software installation.

Some of the harmful programs used in these campaigns are designed to steal information stored in web browsers, including passwords, login sessions, and other account details. Criminals may use this information to access business accounts without permission or impersonate legitimate users. In some cases, stealing a browser session can help attackers bypass additional login protections because the session may already show that the user has successfully signed in. This can put company data, customer information, and internal systems at risk.

These attacks create additional challenges for companies that provide cybersecurity services to other businesses. These providers, known as Managed Security Service Providers (MSSPs), monitor systems, investigate suspicious activity, and help customers respond to threats. When similar scams target multiple customers, security teams may have to investigate the same type of attack repeatedly. This takes time away from other security issues and can delay responses to more serious incidents.

Another challenge is that criminals frequently change the websites, advertisements, and online services they use. A fake website may disappear and be replaced by another one, making it harder for security teams to identify attacks using website addresses alone. Even when the initial warning appears harmless, the activity that follows may reveal a much more serious threat. Security teams therefore need to examine what a suspicious file or link actually does, rather than relying only on how it looks.

To help investigate these threats, cybersecurity teams can use isolated testing environments known as sandboxes. These environments allow analysts to open suspicious files or interact with potentially dangerous websites without exposing a regular computer to the same risks. By watching what happens during the process, investigators can identify harmful commands, suspicious downloads, and attempts to steal information. ANY.RUN says its sandbox technology helps security teams investigate threats more quickly and reports a median threat detection time of 14 seconds. The company also reports 30% fewer escalations from junior analysts to more experienced team members when its sandbox is used.

Businesses can also improve their protection by sharing information about known threats across their security systems. For example, if investigators discover that a particular website or online address is being used to distribute malware, that information can help other security tools recognize related activity. Threat intelligence services collect and organize this type of information so that security teams can identify suspicious websites, files, and other signs of an attack more efficiently. This can help businesses detect related threats before they cause further damage.

Security experts also recommend looking beyond individual incidents to identify patterns that may reveal a wider campaign. If one employee downloads a fake AI application, investigators can search for similar activity elsewhere in the company. They can also check whether other employees visited related websites or ran suspicious commands. Taking this approach can help organizations identify additional affected computers and prevent the same attack from spreading to other parts of the business.

Employees also have an important role to play in preventing these scams. Anyone looking to download an AI application should use the provider’s official website or a trusted application store rather than relying on unfamiliar advertisements or download links. People should be especially cautious when a website asks them to copy and run commands on their computer to complete an installation. Legitimate software should not require users to follow unexpected instructions that they do not understand. If something seems unusual, employees should contact their IT or security team before proceeding.

Businesses should also provide regular training to help employees recognize fake websites, suspicious downloads, and requests for sensitive information. Security teams can strengthen their defenses by monitoring unusual computer activity, keeping software updated, and investigating suspicious messages or files as soon as they are reported. Companies that manage security for multiple customers should also make sure that information discovered during one investigation can be used to protect other customers from similar attacks.

The spread of fake AI scams shows how cybercriminals are adapting their methods to take advantage of the growing popularity of AI tools. By copying trusted brands and disguising malicious activity as ordinary software installation, attackers can trick people into putting their own devices and business accounts at risk. Organizations can reduce this danger by verifying downloads, educating employees, investigating suspicious activity, and sharing threat information across their security systems. As these scams continue to evolve, combining employee awareness with effective security monitoring will be essential to protecting business data and accounts.