Chinese state-affiliated hacking groups are increasing attacks after adding DeepSeek and other open-source artificial intelligence models to their operations, Bloomberg reports, citing researchers at Taiwanese security firm TeamT5. According to Bloomberg, TeamT5 said those groups more than doubled the amount of attacks they carried out after they began assigning routine tasks to AI systems and using the tools to develop more advanced malicious software. The report frames the shift less as a breakthrough in AI-enabled hacking than as a practical operational upgrade: attackers are using readily available models to speed up work that previously consumed human time. The model attribution is not clean. Bloomberg reports that researchers said they could not always determine which AI model was used in a given operation. Still, TeamT5 said DeepSeek’s offerings are popular with hackers in China because of their performance and customizability. The report does not provide a full list of targets, a time window for the measured increase, or raw incident counts in the provided material. What it does establish is that at least one research firm is seeing AI adoption show up in the tempo and tooling of state-linked Chinese cyber activity, including attacks aimed abroad. Who benefits: State-affiliated hacking groups benefit if AI tools reduce the time required for routine operational work. The report also underscores why customizable open-source models are attractive to attackers. Who's exposed: Organizations targeted by Chinese state-affiliated groups abroad are the exposed population identified in the report. The provided material does not specify sectors or countries.