Bing Chat responses infiltrated by ads pushing malware

Not long ago, malicious ads were injected into Microsoft's AI-powered Bing Chat responses, promoting fake download sites that distribute malware.

Microsoft launched Bing Chat, powered by OpenAI's GPT-4 engine, in February 2023 in an attempt to challenge Google's dominance in the search industry.

By providing users with an interactive chat-based experience instead of traditional search queries and result formats, Bing Chat hoped to make online searching more intuitive and user-friendly.

In March, Microsoft began injecting ads into Bing Chat conversations to generate revenue from this new platform.

However, the incorporation of ads into Bing Chat has opened the door for cyber attackers, who are increasingly leveraging search ads to spread malware.

Furthermore, conversing with an AI-powered chat tool may instill unwarranted trust, potentially convincing users to click on ads in a way that they would not do when browsing non-personalized search results.

This conversation-like interaction can lend a false sense of authority and credibility to URLs provided by the AI, so the existing issue of malicious attacks in search platforms is amplified by the introduction of AI assistants.

These ads display promotional tags when a user hovers over a link in a Bing Chat conversation, but this minor measure is too weak to mitigate the risk.

▍Imitating Popular IP Scanners

We discovered malicious ads masquerading as download sites for the popular "Advanced IP Scanner" utility, which has previously been used by RomCom RAT and Somnia ransomware providers.

Researchers found that when you ask Bing Chat how to download Advanced IP Scanner, it displays a download link in the chat.

However, when you hover over the underlined link in the chat, Bing Chat may display the ad first, followed by the legitimate download link. In this case, the sponsored link is a malicious ad pushing malware.

Risky URLs provided in Bing Chat conversations

The operators behind this malicious ad campaign hacked into the ad account of a legitimate Australian business and created two malicious ads targeting system administrators (IP scanner) and lawyers (MyCase legal manager).

Malicious ads created by attackers

Clicking on the IP scanner's malicious ad takes users to a website ("mynetfoldersip[.]cfd") that separates bots and crawlers from human victims by checking various system metrics such as IP address, time zone, and sandbox/virtual machine status.

Victims are then redirected to "advenced-ip-scanner[.]com", a clone of Advanced IP Scanner, which uses a typo (note the 'e' beforehand) to deceive visitors.

Redirection chain

The downloaded MSI installer contains three files, one of which is a heavily obfuscated malicious script that connects to external resources to retrieve the payload.

Malicious VBS present in the downloaded MSI

Unfortunately, the final payload of this malware campaign could not be found, so it remains unclear what malware was ultimately installed.

However, in similar campaigns, attackers typically distribute info-stealing malware or remote access Trojans, allowing them to compromise other accounts or corporate networks.

Malicious ads in Bing Chat conversations highlight the expanding frontier of cyber threats; users must remain vigilant about chatbot results and double-check URLs before downloading anything.

For enterprises, SecurityVerify recommends taking proactive defensive measures, conducting security verification before attacks occur to identify weaknesses in the security defense system and optimize and upgrade them in a timely manner.

The SecurityVerify simulated attack library collects various attack methods from classic as well as currently popular and active threat groups. You can select or orchestrate corresponding attack scenarios in the SecurityVerify Security Metric Verification Platform to verify with one click whether your security defense system can respond effectively.