ChatGPT Grapples with Consecutive Days of Service Disruptions

San Francisco, CA – The widely used artificial intelligence chatbot ChatGPT has once again experienced significant service disruptions, marking the second consecutive day users have reported issues. OpenAI, the company behind the advanced AI model, has confirmed ongoing problems, though it states that resolutions are being rolled out.

Reports of widespread problems began surfacing on Wednesday morning, with users unable to log in or engage in conversations with the AI. The outage monitoring website Downdetector saw a substantial spike in user-submitted issues, with one report indicating over 15,000 users flagged problems around 12:30 p.m. ET on Wednesday. This follows a similar pattern of disruptions experienced just the day prior.

According to information released by OpenAI, at least two distinct issues were identified. The first, reported approximately 20 minutes before the initial publication of a live blog tracking the situation, was described as "Elevated error rates for finetuning jobs." Shortly after, about 44 minutes prior to that update, a second issue was confirmed: "Elevated errors for ChatGPT and Platform users." Both of these could manifest as the chatbot failing to respond to queries or experiencing significant delays in its responses.

While ChatGPT itself, when queried directly about its status, reportedly indicated it was functioning normally, its parent company, OpenAI, has acknowledged the widespread nature of the problems. The company stated on Wednesday morning that "custom GPT updates are failing for users," but that a fix was implemented by 1 p.m. Eastern time. An update from OpenAI read, "We have applied the mitigation and are monitoring the recovery."

The timing of these disruptions is particularly noteworthy, as users are increasingly relying on AI tools for a variety of tasks, from creative writing and coding assistance to research and general information gathering. The intermittent availability of such a prominent service can cause significant workflow interruptions for individuals and businesses alike.

TechRadar reported that the number of reported issues on Downdetector peaked at over 12,000 before reportedly dropping to under 300 in the United States, suggesting a partial or full restoration of service. However, the recurrence of these problems on consecutive days raises questions about the underlying stability of OpenAI's infrastructure and its ability to handle peak demand or unexpected technical glitches.

As of the latest updates, OpenAI has indicated that fixes have been deployed for both the finetuning job errors and the general ChatGPT and Platform user errors. The company continues to monitor the recovery process. The incident underscores the ongoing challenges in maintaining the reliability of complex AI systems that are becoming integral to daily digital life.

This latest series of outages follows a pattern of intermittent issues that have affected various AI platforms. While the exact cause of the consecutive disruptions remains under investigation by OpenAI, the company's swift confirmation and stated efforts to implement fixes are crucial for regaining user trust and ensuring the continued accessibility of their flagship AI product.