Several of the world’s leading artificial intelligence platforms experienced significant service disruptions on Thursday, September 3, affecting thousands of users globally.

OpenAI’s ChatGPT, Google Gemini and Anthropic’s Claude were among the major AI services affected, according to reports recorded by outage-tracking services.

The disruption also appeared to affect Grok and Amazon Web Services.

Reports began increasing sharply during the evening in Indian Standard Time, with users describing problems ranging from websites becoming inaccessible to internal server errors while attempting to generate AI responses.

CHATGPT AND CODEX AFFECTED

Outage reports involving OpenAI rose sharply during the disruption.

Users reported difficulties accessing both the ChatGPT web interface and mobile application, while others encountered errors when trying to submit prompts or generate responses.

OpenAI acknowledged problems affecting its services through its status page, reporting elevated errors across ChatGPT and Codex.

The company subsequently said it had applied mitigation measures and was monitoring the recovery.

GEMINI AND CLAUDE USERS REPORT ERRORS

Google Gemini users also reported service problems during the wider disruption.

Some users were unable to complete prompts, while reports also described timeout-related errors.

Anthropic’s Claude experienced similar difficulties.

Anthropic said it had identified the cause of elevated errors affecting requests to several Claude models and was working on a fix.

The company later reported that most affected models had returned to their baseline error rate, while Opus 4.8 and Opus 5 remained affected at that stage of the recovery.

MULTIPLE MAJOR SERVICES HIT

The near-simultaneous problems across several major AI platforms drew attention because millions of consumers, developers and businesses increasingly depend on generative AI services for everyday tasks.

ChatGPT, Gemini and Claude are used for activities ranging from writing and research assistance to software development and enterprise workflows.

Even relatively short periods of disruption can therefore affect a large number of individual and business users.

The incident also highlighted the growing dependence on cloud infrastructure and large-scale computing systems behind generative AI services.

As providers worked to restore their platforms, service status information indicated that recovery was progressing, although the timing and extent of disruption varied between platforms.