DeepSeek is becoming more than an AI tool for coding, research and everyday tasks. Security experts found that hackers are also using the Chinese AI model to support cyberattacks. Its ability to write code, explain technical steps and automate parts of a task can make attacks quicker and easier to run.
The growing concern is not that DeepSeek can suddenly carry out an entire cyberattack on its own. The bigger issue is how it can help attackers with jobs that once needed more time and skill. A recent report by TeamT5 brought this to light. Charles Li, chief analyst at Team T5, stated, “DeepSeek is the AI of choice for Chinese hackers because it’s relatively powerful with very low cyber guardrails.” Further, he added, “Western models are highly sought-after but their guardrails are much stricter and require a lot more effort to bypass.”
Hackers can use DeepSeek to break into someone’s email system and analyze technical information to solve problems faster. This can lower the effort needed to launch certain attacks and allow criminals to focus on other parts of their campaigns.
Automation is changing the cost of cybercrime. A task that once took hours of manual work can now be completed much faster with the help of an AI model. This can be useful for attackers who want to test code, create fake messages or find ways around basic security systems.
The biggest risk is scale. When AI handles smaller tasks, attackers can repeat them across more targets without needing a large team of skilled people. This could make common cyber threats more frequent.
DeepSeek is also part of a much wider trend. Hackers have been exploring several AI models to improve their work. The technology itself is not the problem. The concern comes from how quickly criminals are learning to use it.
AI will likely remain a double-edged sword for cybersecurity. Companies can use it to spot threats, while attackers can use similar tools to improve their methods.
As AI becomes easier to access, security teams may need to prepare for attacks that are faster, cheaper and harder to predict. DeepSeek's growing use in cybercrime is another sign that the AI race is no longer limited to legitimate users.