Man vs machine: OpenAI’s ChatGPT grapples with potential DDoS attack

Artificial intelligence startup OpenAI said on Wednesday it is dealing with “abnormal traffic patterns” on its ChatGPT platform suggests hackers are trying to overwhelm its services with a distributed denial-of-service (DDoS) attack.

OpenAI, which helped popularize generative AI with its ChatGPT chatbot, said in a system update that it spotted signs of a DDoS attack, where external perpetrators attempt to overload a platform by pinging it repeatedly.

“We are dealing with periodic outages due to an abnormal traffic pattern reflective of a DDoS attack. We are continuing work to mitigate this,” the San Francisco-based startup told media.

The company’s latest update follows what it termed a “major outage” across ChatGPT and developer tools that triggered high error rates. OpenAI said it had fixed that issue, which it initially blamed on unusually high demand rather than a malicious attack.

OpenAI held its first developer conference on Monday, where it introduced a preview of GPT-4 Turbo, a faster version of its latest natural language model that powers ChatGPT.

ChatGPT can now access real-time information

Roughly 100 million people now use ChatGPT weekly, OpenAI said at the conference, and more than 90% of Fortune 500 companies are building tools using its platform.

With the popularity of ChatGPT, users have been searching for answers on why ChatGPT is not working during the outages and attempting to log in repeatedly. The issues have also led some to explore ChatGPT alternatives and paraphrasing tools.

OpenAI faces tough rivalry from Google’s Bard, Antropic’s Claude 2 AI system and Grock – a well-funded AI startup by billionaire Elon Musk