The disruption of Microsoft services, such as Outlook, Teams, and CoPilot, extended to working hours in the Asia-Pacific region, today, Friday, after affecting thousands of users in the United States, while the technology giant faced difficulties in fixing server problems that it blamed for the outage.
The outage affected government agencies in Australia, Japan and the United States, as well as companies including OpenAI and Anthropic. This is the second outage for Microsoft 365 services this year, raising concerns that recurring outages will lead business customers to question Microsoft's reliability.
Partial improvement in Microsoft services
Microsoft said in its latest statement that it had made modifications to the affected servers, but this step did not succeed in returning the service to normal in a stable manner. It indicated partial improvement in the areas where these amendments were implemented.
The company said on the “X” platform: “We are reviewing additional measures that we can take to improve the stability of the infrastructure in general.”
DownDetector, which tracks outages by compiling status reports from a number of sources, reported more than 18,000 reports of issues with Outlook in the United States at 11:30 p.m. EST (0430 GMT). It also observed outages in a number of Microsoft services in the Asia-Pacific region, including India, Australia, and Japan, early today, Friday, local time.
At the height of the holiday, Down Detector detected more than 130,000 reports of problems with Outlook in the United States at around 1 p.m. EST, Thursday. The malfunction lasted for about 12 hours.
Microsoft said in a notice about the service status that features of Microsoft 365 services that rely on Online Exchange and email were affected, while elements that do not depend on Online Exchange remained functional. Exchange Online is the company's cloud messaging platform, which provides email, calendar, contacts and tasks services.
The company added that it “does not have any information” indicating that the malfunction was the result of malicious electronic activity or a cyber attack.
Australian Home Affairs and Cyber Security Minister Tony Burke said that Microsoft does not believe it is dealing with a cyber attack.
“They believe they are dealing, in some ways, with something similar to what we saw with CrowdStrike, where a software update led to a glitch,” Burke explained in a press conference.
“The impact they are dealing with... is largely focused on email systems,” he added.
The holidays extend to government agencies, airlines, and artificial intelligence
The Australian Intelligence Agency and the Federal Police were among those affected by this outage on Friday.
The 2024 Crowd Strike outage was one of the largest Internet disruptions, crippling technology systems in hospitals, banks and airports, highlighting the fragility of interconnected technologies around the world.
The US House of Representatives indicated that the malfunction affected its users, and that its engineers were working with Microsoft.
Japanese Deputy Chief Cabinet Secretary Masanao Ozaki told reporters that the malfunction affects the infrastructure of government information systems.
The two American artificial intelligence companies, “OpenAI” and “Anthropic,” said that they encountered problems with Microsoft’s connectivity services, while the Indian airlines, “IndiGo” and “Acasa Air,” said that their email support services were affected.
Josh Gilbert, a market analyst at the eToro trading platform, pointed out that investors are likely to overcome the impact of the outage, but repeated outages may harm Microsoft in the long term.
Microsoft faced a disruption in its 365 services last January, as well as in October of last year.
Last month, Microsoft reported a malfunction in its cloud platform (Azure), affecting its productivity software package and a group of industries around the world.
A group of non-profit organizations filed a lawsuit today, Thursday, against the companies “OpenAI”, “Anthropic”, “MetaPlatforms”, “Google” and the artificial intelligence division of “SpaceX”, seeking to stop the development of the most advanced artificial intelligence models, based on concerns expressed by the leaders of these companies themselves that this technology poses a threat to humanity.
These organizations, which focus on accountability in technology, filed the lawsuit in San Francisco. She said that California's laws against unfair business practices provide the basis for issuing an injunction preventing companies from continuing to develop products, such as “ChatGPT,” “Gemini,” and “Cloud.”
In the lawsuit, the organizations cite statements by leaders, such as Dario Amodei, CEO of Anthropic, and Elon Musk, CEO of SpaceX, about the ability of artificial intelligence to destroy civilizations. The lawsuit also refers to recent incidents in which models went out of testing environments and automatically attacked government websites.
The text of the lawsuit stated, according to Reuters: “No other sector is allowed to consider the development of unsafe products, let alone those that could kill us all, as inevitable... Nor is any sector given the right to regulate on its own the development of such dangerous products.”
In the lawsuit, the organizations accuse the companies of engaging in business practices that cause public harm or harm public health and the well-being of society, pointing to past incidents that have already occurred and the possibility of future harm.
A Meta company spokesman declined to comment. Representatives of the other companies did not respond to requests for comment.
Other lawsuits
US President Donald Trump's administration has so far refused to seek widespread regulation of artificial intelligence, fearing that this would harm the United States' ability to compete with China, but states and private entities have sought to use lawsuits to force companies to make changes.
Florida Attorney General James Othmeyer filed a lawsuit against OpenAI in June. He accused the company of misleading the public about the safety of its “GBT Chat” platform, and said it had harmed children by providing perpetrators of school shootings with information and guidance on self-harm, in addition to causing addiction among young users.
The state of Florida asked the judge hearing the case in September to prevent OpenAI from developing new artificial intelligence models without external oversight. That request is still pending, according to court records.
After the state filed the lawsuit, Drew Pusateri, a spokesman for OpenAI, said the company had temporarily stopped training its most efficient models and would not resume until additional safeguards were in place.
A group of eight states filed other lawsuits against OpenAI last month in various courts around the country. She said that GPT chat directed its users, including teenagers, to harm themselves or others. OpenAI did not respond to requests for comment after the lawsuits were filed.
Through their lawsuits, the states seek to impose financial penalties and issue an order preventing the company from developing self-developing artificial intelligence models unless it can ensure that those models are compatible with human values.
DownDetector, a website that specializes in tracking Internet outages, showed that Microsoft 365, Outlook, and Teams services stopped working for thousands of users in the United States on Thursday, and this outage affected the connectivity connectors used by artificial intelligence services provided by OpenAI and Anthropic.
DownDetector, which tracks outages by compiling status reports from a number of sources, reported more than 97,000 instances of users reporting problems with Outlook at 2:07 p.m. EST. The actual number of affected users may differ from what is reported by the site.
Down Detector recorded more than 38,000 reports from users regarding problems with the Microsoft 365 service, and about 4,500 reports from users facing difficulty accessing the Teams application.
Microsoft said in a statement: “We are doing our best to limit the impact of these problems.”
She stated that her team has stopped all deployment operations, is working to redirect additional data traffic away from the affected infrastructure, and is restarting the affected servers to return them to service.
The company said earlier today that it had fixed an outage in the Teams service that affected users in North America, and that the problem had been completely resolved.
The company later stated that the issue may also affect Microsoft 365 service for users in other regions.
OpenAI said it was investigating an issue with Microsoft's connectors, where new connections may break and the tools may not work as expected.
Anthropic reported that it had detected a problem with one of its service providers, affecting its connection to Microsoft 365.
