AI Crisis

Following a prolonged outage during the early morning hours, OpenAI’s ChatGPT chatbot experienced another interruption, joined by other AI providers. On Tuesday morning, both Anthropic’s Claude and Perplexity encountered issues, though they were swiftly resolved compared to ChatGPT.

While Google’s Gemini seems to be operating well, it also encountered a brief outage, according to some user reports. The simultaneous downtime of three major AI providers raises concerns about a potential broader infrastructure or internet-scale problem. This situation is reminiscent of the issues that have simultaneously impacted multiple social media platforms, underscoring the potential implications of these outages.

Furthermore, the disruptions experienced by Claude and Perplexity might not solely stem from bugs or technical glitches but could also result from an influx of traffic triggered by ChatGPT’s outage.

ChatGPT’s landing page was non-functional during the outage, preventing users from messaging the AI chatbot. The issue began around 7:33 AM PT and was resolved by approximately 10:17 AM PT, marking another multi-hour disruption.

ChatGPT Offline
Image Credits: OpenAI/ChatGPT landing page down

At a particular time during the outage, ChatGPT’s website was updated with a message indicating that the service had reached capacity, expressed in pirate speak. Users were then allowed to receive notifications once the service was restored.

ChatGPT Capacity
Image Credits: OpenAI/capacity message

Claude’s website displayed an error message stating, “An error occurred in the Server Components render. The specific message is omitted in production builds to avoid leaking sensitive details. A digest property is included on this error instance which may provide additional details about the nature of the error.”

The site then advises to “try again.” 

Claude started working again sometime after 12:10 p.m. ET.

Anthropic Claude down
Image Credits: Anthropic/Claude down

Perplexity’s website also displayed an over-capacity message, which is typically displayed when the service receives an excessive number of requests. The message read: “We’ll be right back” and “We’re getting a lot of questions right now and have reached our capacity. Please come back soon.”

The site resumed functioning when Claude’s outage was resolved but has experienced intermittent ups and downs since then.

Perplexity outrage
Image Credits: Perplexity outrage

One thing worth noting is that ChatGPT had only recently resolved a lengthy outage that affected its service overnight. The chatbot was down for users worldwide during the early morning hours (Pacific Time), leading to complaints across X and Threads. The company’s status page indicated that a bug was being fixed, and the issue was being monitored around 12:21 a.m. PT.

By 7:33 a.m. PT, the status page was updated to report that ChatGPT was once again unavailable for some users, with OpenAI “currently investigating this issue.” The chatbot remained down as of 12:28 p.m. ET, with the status page flagging it as a “major outage.” The issues were resolved several hours later.

ChatGPT Status
Image Credits: OpenAI status page
https://t.co/dUkVhSehGG

The AI providers did not immediately respond to requests for comment or provide additional information on the outages.

https://t.co/06PIS0lAv3

In addition to the AI outages, other online disruptions are occurring at present. Forbes reports a zero-day spam attack on TikTok, which affected select celebrity and brand accounts, such as Paris Hilton and CNN.