Chinese AI Company DeepSeek Recovering From Apparent DDoS Cyber Attack

by | Feb 4, 2025

The first Chinese AI company to contend with the big Western players has run into some trouble in its first month, not the least of which was an apparent cyber attack that shut down new DeepSeek registrations for a time.

The company seems to have almost fully recovered at this point, but the cyber attack is part of a string of incidents that have raised security and safety questions. DeepSeek rocketed to the top of the app download charts after being released on January 11, but its seeming capability is offset by the fact that little is really still known about the company and its inner workings.

Chinese AI company’s rise to stardom already full of complications

The emergence of DeepSeek over the past month has been compared to the launch of Sputnik in 1957, announcing that the field is now a neck-and-neck technological race between China and the US. But the now-prominent Chinese AI company has been beset with issues in its first few weeks of relevance: jailbreaking, censorship, hallucination issues, and the accidental exposure of private conversations among others.

The DDoS cyber attack is among the more serious of these early issues, if only because it has raised serious questions about the Chinese AI company’s capacity for traffic. The service had already been struggling to some degree before the attack started, with users noting that it was more frequently asking them to try conversations again later due to “heavy traffic” as it quickly tallied more and more app downloads. After the attack began the response was to shut down new registrations for several days, raising some questions as to whether the whole thing was just cover for a scaling issue.

For a time after the cyber attack was announced, only users in mainland China were able to register new accounts using a local phone number. The Chinese AI company has indicated that it now has a handle on whatever the issue was and is implementing a fix.

Cyber attack follows exposure of records, accusations of theft

In late January DeepSeek became the most-downloaded Apple app, but for some time new users were likely frustrated by inability to set up a new account. Pre-existing users that had registered prior to the cyber attack do not appear to have experienced any disruption.

A formal statement from the Chinese AI company could clear up some of the confusion about what exactly happened, and is virtually a necessity at this point given the other issues it has experienced within a short amount of time. DeepSeek has been peppered with accusations of stealing the work of OpenAI, had security researchers trash it as being fundamentally more unsafe than its more established rivals, and experienced a breach in which private conversations and other internal materials appear to have been left open to the internet. It also appears to do PR work for the Chinese government, avoiding certain uncomfortable questions or answering them with CCP talking points.

But regardless of these safety and security issues, DeepSeek has already changed the paradigm and shifted the flow of AI investment money. It appears that AI capability comparable to ChatGPT and its rivals can be achieved without the latest and greatest chips, and without the same quantity of data use, which means less money can be spent on it. The big question is whether its open-source model, key to much of this cost-cutting and rapid innovation, can be proven secure enough to build future projects on.

Recent Posts

How can we help?

13 + 5 =

× How can I help you?