Vietnam.vn - Nền tảng quảng bá Việt Nam

ChatGPT and Sora Global Outbreak: A Warning from an AI-Dependent World

On June 10, a major technical issue caused ChatGPT and Sora to stop working, seriously affecting users. Let's find out the cause and lessons from this incident.

Báo Tuổi TrẻBáo Tuổi Trẻ11/06/2025

ChatGPT - Ảnh 1.

ChatGPT and Sora 'lost connection' across the board

June 10 marked one of the most serious technical issues OpenAI has ever encountered: its ChatGPT platform and Sora video creation tool simultaneously went down for hours, affecting tens of millions of users worldwide. Individual users, small businesses, and corporations integrating AI APIs were all affected.

OpenAI responded early, but not fast enough

According to Tuoi Tre Online 's investigation, the incident started at around 2:45 p.m. Vietnam time. Immediately, the number of error reports skyrocketed on incident monitoring platforms such as DownDetector, with more than 1,800 reports in just 30 minutes.

Users around the world from the US, UK, Germany, Japan, India to Vietnam have reported being unable to access ChatGPT, API not responding, or user interface errors. As for Sora - OpenAI's artificial intelligence video creation platform, the entire service is almost paralyzed.

Within the first 30 minutes, OpenAI updated its system status page with a note: "We are investigating a technical issue affecting ChatGPT and related services." However, it was not until nearly 5 hours later that the company confirmed that it had "identified the root cause and is working on a fix."

Full service restoration only occurred around 6:30 a.m. the following morning Vietnam time - nearly 16 hours after the error first appeared. While the API and most of ChatGPT's core functionality were back to normal, some advanced features such as voice mode continued to experience intermittent issues.

What's the reason? OpenAI hasn't released details yet.

ChatGPT - Ảnh 2.

The specific cause has not yet been announced by OpenAI.

While OpenAI did not disclose the specific cause, the company confirmed that the issue originated from the system's core infrastructure.

Meanwhile, technology experts have put forward some interesting hypotheses. One possibility is that the system was overloaded or had a load balancing error due to the surge in users - ChatGPT is currently estimated to serve more than 400 million users per week.

Additionally, a technical error arising from an unstable internal update could also be a potential cause.

Some opinions also question the possibility of the platform being attacked by a distributed denial of service (DDoS), although there is currently no official confirmation from OpenAI regarding this hypothesis.

User reactions and lessons on AI dependence

On social network X (formerly Twitter), the hashtag #ChatGPTDown quickly climbed to the top of the trend. Many people expressed confusion at not being able to use the tool that had become their daily "digital assistant". Some happily shared that they "had to use their real memory" to write emails, come up with ideas or look up information.

But for businesses and developers integrating AI APIs into their systems, the fallout has been far from benign. Many chatbot websites have been disabled, data analytics tools have been disrupted, and automated customer service systems have been down for hours.

'This incident is a clear warning about the increasing dependence on AI in business operations and daily life,' a technology expert told Tuoi Tre Online.

OpenAI faces the challenge of operating at global scale

ChatGPT - Ảnh 3.

Most OpenAI services are now back to normal.

Since late 2023, OpenAI has continuously launched products such as GPT-4.5 Turbo, Sora, ChatGPT Voice... helping the company become the center of the commercial AI ecosystem. However, this comes with huge pressure on technical infrastructure.

“A single second of AI delay can lead to a chain of operational delays,” MarketWatch analyzed. The lesson here is to build in backup solutions, both on the user and the provider side.

As of June 11, most of OpenAI’s services are back up and running. However, some features, such as voice mode, are still experiencing sporadic issues. OpenAI says it is continuing to monitor the system and will continue to improve stability in the coming days.

Recovered, but risks remain

This incident is not just a simple technical glitch, but a warning bell for the entire technology world: AI is powerful but not invincible.

The design of technical systems and social infrastructure must be accompanied by contingency thinking, risk calculation and self-reliance when necessary.

Back to topic
PHAN HAI DANG

Source: https://tuoitre.vn/su-co-toan-cau-cua-chatgpt-va-sora-canh-bao-tu-mot-the-gioi-phu-thuoc-ai-20250611102155514.htm


Comment (0)

No data
No data

Heritage

Figure

Enterprise

No videos available

News

Political System

Destination

Product