ChatGPT users have been having issues using the AI chatbot over the last day or so. The service from OpenAI has had a number of periodic outages during that time, and it looks like it might be due to a denial of service attack or DDoS.
CNBC reports that the first outages started on Wednesday around 9 a.m. Eastern time. About an hour later, OpenAI posted word on its status page that it had fixed the problem. However, the status page confirmed later in the day that more outages and issues were affecting ChatGPT and its API.
As of this writing, the status page says that OpenAI’s API and ChatGPT services are currently experiencing “Degraded Performance.” The last message on the state page, posted on Wednesday evening, stated the following:
We are dealing with periodic outages due to an abnormal traffic pattern reflective of a DDoS attack. We are continuing work to mitigate this.
CNBC also reported that another AI chatbot, Anthropic’s Claude 2, had some downtime on Wednesday as well. It’s unknown if the two events are related.
This possible DDoS attack occurred just after OpenAI held its first developer’s conference, DevDay, on Tuesday. Among other things, the company announced a new AI product, GPTs, which will allow developers and companies to customize their own versions of ChatGPT to serve their specific needs and services.
It also announced a preview of its GPT-4 Turbo model. It will have a 128k context window, which OpenAI says will allow it to accept a single text prompt equivalent to 300 pages. It will also work faster than the current GPT-4 chatbot while simultaneously costing up to two times less to operate. A new version of GPT-3.5 Turbo, which will support a 16k context window by default, was also announced by OpenAI this week.