Ki-powered Cyber ​​crimes with the risk of exceeding traditional immune system warns Check Point.

Tave speak:

• Hackers are increasingly using AI in their attacks and the defenders should follow Sucheck Point software technologies, which shared in a report published on Wednesday.
• The company's AI security report, which was announced at the RSAC conference 2025 in San Francisco, also found that one of 13 generative AI inputs potentially confidential information, and one of 80 inputs issued “a high risk of sensitive data injury”.
• According to Check Point, unauthorized AI tools, data loss and weaknesses of the AI ​​platform stated the list of AI risks for companies.

Diver Insight:

Check Point's report is a roadmap for companies that are interested in the development of AI tools for the rationalization of their processes. It also offers companies warnings for companies that are not interested in using AI – namely that poor access controls and software management enable their employees to use AI tools without permission.

Whether authorized or not authorized, AI tools can reveal corporate secrets, user data and others Sensitive information. The Check Point Report cataloged the risks that companies have to be aware of. For example, it warns that non -approved AI software “can lead to security gaps, compliance problems and inconsistent data management and a company is exposed to a company risk or data injuries”. With regard to data loss, the report calls on to evaluate “AI applications for data protection and best practice in the industry” before they are accepted, and find that some applications store data, share them with third parties or do not protect them from hackers.

The report also offers a sample of the findings of the researchers of the Check Point researchers about Hacker Use of AI. “Chatgpt and Openais API are currently the most popular models for cybercriminals, while others like Google Gemini, Microsoft Copilot and Anthropic Claude quickly gain popularity,” says the report, referring to information on the dark web.

At the same time, according to Check Point, the AI ​​software “Landscape with the start of open source models such as Deepseek and Qwen from Alibaba” changes, which “only have minimal restrictions on use”. In addition, the rise of tailor-made AI programs such as Wormgpt and the FunkSec Ransomware Group is AI-operated DDOS tool – is an important development in the cyber criminal ecosystem that cyber defenders should be careful.

The report by Check Point makes the case that the defenders should involve their defense strategies with the inclusion of the AI ​​also in the form of their own anomaly recognition platform of corporate in order to counteract the accelerating effects of the offensive AI. “The use of AI in cybercrime is no longer theoretical,” wrote the researchers. “It develops parallel to the mainstream -Ki introduction and in many cases moves faster than conventional security controls adapt.