SolarWinds Hackers Accessed Microsoft Source Code, the Company Says

The Microsoft logo is pictured ahead of the Mobile World Congress in Barcelona, Spain February 24, 2019. (Reuters)
The Microsoft logo is pictured ahead of the Mobile World Congress in Barcelona, Spain February 24, 2019. (Reuters)
TT

SolarWinds Hackers Accessed Microsoft Source Code, the Company Says

The Microsoft logo is pictured ahead of the Mobile World Congress in Barcelona, Spain February 24, 2019. (Reuters)
The Microsoft logo is pictured ahead of the Mobile World Congress in Barcelona, Spain February 24, 2019. (Reuters)

The hacking group behind the SolarWinds compromise was able to break into Microsoft Corp and access some of its source code, Microsoft said on Thursday, something experts said sent a worrying signal about the spies’ ambition.

Source code - the underlying set of instructions that run a piece of software or operating system - is typically among a technology company’s most closely guarded secrets and Microsoft has historically been particularly careful about protecting it.

It is not clear how much or what parts of Microsoft’s source code repositories the hackers were able to access, but the disclosure suggests that the hackers who used software company SolarWinds as a springboard to break into sensitive US government networks also had an interest in discovering the inner workings of Microsoft products as well.

Microsoft had already disclosed that like other firms it found malicious versions of SolarWinds’ software inside its network, but the source code disclosure - made in a blog post - is new. After Reuters reported it was breached two weeks ago, Microsoft said it had not “found any evidence of access to production services.”

Three people briefed on the matter said Microsoft had known for days that the source code had been accessed. A Microsoft spokesman said security employees had been working “around the clock” and that “when there is actionable information to share, they have published and shared it.”

The SolarWinds hack is among the most ambitious cyber operations ever disclosed, compromising at least half-a-dozen federal agencies and potentially thousands of companies and other institutions. US and private sector investigators have spent the holidays combing through logs to try to understand whether their data has been stolen or modified.

Modifying source code - which Microsoft said the hackers did not do - could have potentially disastrous consequences given the ubiquity of Microsoft products, which include the Office productivity suite and the Windows operating system. But experts said that even just being able to review the code could offer hackers insight that might help them subvert Microsoft products or services.

“The source code is the architectural blueprint of how the software is built,” said Andrew Fife of Israel-based Cycode, a source code protection company.

“If you have the blueprint, it’s far easier to engineer attacks.”

Matt Tait, an independent cybersecurity researcher, agreed that the source code could be used as a roadmap to help hack Microsoft products, but he also cautioned that elements of the company’s source code were already widely shared - for example with foreign governments. He said he doubted that Microsoft had made the common mistake of leaving cryptographic keys or passwords in the code.

“It’s not going to affect the security of their customers, at least not substantially,” Tait said.

Microsoft noted that it allows broad internal access to its code, and former employees agreed that it is more open than other companies.

In its blog post, Microsoft said it had found no evidence of access “to production services or customer data.”

“The investigation, which is ongoing, has also found no indications that our systems were used to attack others,” it said.

Reuters reported a week ago that Microsoft-authorized resellers were hacked and their access to productivity programs inside targets leveraged in attempts to read email. Microsoft acknowledged some vendor access was misused but has not said how many resellers or customers may have been breached.

There was no response to requests for comment from the FBI, which is investigating the hacking campaign, or from the Department of Homeland Security’s Cybsersecurity and Infrastructure Security Agency.

US officials have attributed the SolarWinds hacking campaign to Russia, an allegation the Kremlin denies.

Both Tait and Ronen Slavin, Cycode’s chief technology officer, said a key unanswered question was which source code repositories were accessed. Microsoft has a huge range of products, from widely used Windows to lesser known software such as social networking app Yammer and the design app Sway.

Slavin said he was worried by the possibility that the SolarWinds hackers were poring over Microsoft’s source code as prelude to a much more ambitious offensive.

“To me the biggest question is, ‘Was this recon for the next big operation?’” he said.



AI No Better Than Other Methods for Patients Seeking Medical Advice, Study Shows

AI (Artificial Intelligence) letters and a robot hand are placed on a computer motherboard in this illustration created on June 23, 2023. (Reuters)
AI (Artificial Intelligence) letters and a robot hand are placed on a computer motherboard in this illustration created on June 23, 2023. (Reuters)
TT

AI No Better Than Other Methods for Patients Seeking Medical Advice, Study Shows

AI (Artificial Intelligence) letters and a robot hand are placed on a computer motherboard in this illustration created on June 23, 2023. (Reuters)
AI (Artificial Intelligence) letters and a robot hand are placed on a computer motherboard in this illustration created on June 23, 2023. (Reuters)

Asking AI about medical symptoms does not help patients make better decisions about their health than other methods, such as a standard internet search, according to a new study published in Nature Medicine.

The authors said the study was important as people were increasingly turning to AI and chatbots for advice on their health, but without evidence that this was necessarily the best and safest approach.

Researchers led by the University of Oxford’s Internet Institute worked alongside a group of doctors to draw up 10 different medical scenarios, ranging from a common cold to a life-threatening hemorrhage causing bleeding on the brain.

When tested without human participants, three large-language models – Open AI's Chat GPT-4o, ‌Meta's Llama ‌3 and Cohere's Command R+ – identified the conditions in ‌94.9% ⁠of cases, ‌and chose the correct course of action, like calling an ambulance or going to the doctor, in an average of 56.3% of cases. The companies did not respond to requests for comment.

'HUGE GAP' BETWEEN AI'S POTENTIAL AND ACTUAL PERFORMANCE

The researchers then recruited 1,298 participants in Britain to either use AI, or their usual resources like an internet search, or their experience, or the National Health Service website to ⁠investigate the symptoms and decide their next step.

When the participants did this, relevant conditions were identified in ‌less than 34.5% of cases, and the right ‍course of action was given in ‍less than 44.2%, no better than the control group using more traditional ‍tools.

Adam Mahdi, co-author of the paper and associate professor at Oxford, said the study showed the “huge gap” between the potential of AI and the pitfalls when it was used by people.

“The knowledge may be in those bots; however, this knowledge doesn’t always translate when interacting with humans,” he said, meaning that more work was needed to identify why this was happening.

HUMANS OFTEN GIVING INCOMPLETE INFORMATION

The ⁠team studied around 30 of the interactions in detail, and concluded that often humans were providing incomplete or wrong information, but the LLMs were also sometimes generating misleading or incorrect responses.

For example, one patient reporting the symptoms of a subarachnoid hemorrhage – a life-threatening condition causing bleeding on the brain – was correctly told by AI to go to hospital after describing a stiff neck, light sensitivity and the "worst headache ever". The other described the same symptoms but a "terrible" headache, and was told to lie down in a darkened room.

The team now plans a similar study in different countries and languages, and over time, to test if that impacts AI’s performance.

The ‌study was supported by the data company Prolific, the German non-profit Dieter Schwarz Stiftung, and the UK and US governments.


Meta Criticizes EU Antitrust Move Against WhatsApp Block on AI Rivals

(FILES) This illustration photograph taken on December 1, 2025, shows the logo of WhatsApp displayed on a smartphone's screen, in Frankfurt am Main, western Germany. (Photo by Kirill KUDRYAVTSEV / AFP)
(FILES) This illustration photograph taken on December 1, 2025, shows the logo of WhatsApp displayed on a smartphone's screen, in Frankfurt am Main, western Germany. (Photo by Kirill KUDRYAVTSEV / AFP)
TT

Meta Criticizes EU Antitrust Move Against WhatsApp Block on AI Rivals

(FILES) This illustration photograph taken on December 1, 2025, shows the logo of WhatsApp displayed on a smartphone's screen, in Frankfurt am Main, western Germany. (Photo by Kirill KUDRYAVTSEV / AFP)
(FILES) This illustration photograph taken on December 1, 2025, shows the logo of WhatsApp displayed on a smartphone's screen, in Frankfurt am Main, western Germany. (Photo by Kirill KUDRYAVTSEV / AFP)

Meta Platforms on Monday criticized EU regulators after they charged the US tech giant with breaching antitrust rules and threaten to halt its block on ⁠AI rivals on its messaging service WhatsApp.

"The facts are that there is no reason for ⁠the EU to intervene in the WhatsApp Business API. There are many AI options and people can use them from app stores, operating systems, devices, websites, and ⁠industry partnerships," a Meta spokesperson said in an email.

"The Commission's logic incorrectly assumes the WhatsApp Business API is a key distribution channel for these chatbots."


Chinese Robot Makers Ready for Lunar New Year Entertainment Spotlight

A folk performer breathes fire during a performance ahead of Lunar New Year celebrations in a village in Huai'an, in China's eastern Jiangsu Province on February 7, 2026. (AFP)
A folk performer breathes fire during a performance ahead of Lunar New Year celebrations in a village in Huai'an, in China's eastern Jiangsu Province on February 7, 2026. (AFP)
TT

Chinese Robot Makers Ready for Lunar New Year Entertainment Spotlight

A folk performer breathes fire during a performance ahead of Lunar New Year celebrations in a village in Huai'an, in China's eastern Jiangsu Province on February 7, 2026. (AFP)
A folk performer breathes fire during a performance ahead of Lunar New Year celebrations in a village in Huai'an, in China's eastern Jiangsu Province on February 7, 2026. (AFP)

In China, humanoid robots are serving as Lunar New Year entertainment, with their manufacturers pitching their song-and-dance skills to the general public as well as potential customers, investors and government officials.

On Sunday, Shanghai-based robotics start-up Agibot live-streamed an almost hour-long variety show featuring its robots dancing, performing acrobatics and magic, lip-syncing ballads and performing in comedy sketches. Other Agibot humanoid robots waved from an audience section.

An estimated 1.4 million people watched on the Chinese streaming platform Douyin. Agibot, which called the promotional stunt "the world's first robot-powered gala," did not have an immediate estimate for total viewership.

The ‌show ran a ‌week ahead of China's annual Spring Festival gala ‌to ⁠be aired ‌by state television, an event that has become an important - if unlikely - venue for Chinese robot makers to show off their success.

A squad of 16 full-size humanoids from Unitree joined human dancers in performing at China Central Television's 2025 gala, drawing stunned accolades from millions of viewers.

Less than three weeks later, Unitree's founder was invited to a high-profile symposium chaired by Chinese President Xi Jinping. The Hangzhou-based robotics ⁠firm has since been preparing for a potential initial public offering.

This year's CCTV gala will include ‌participation by four humanoid robot startups, Unitree, Galbot, Noetix ‍and MagicLab, the companies and broadcaster ‍have said.

Agibot's gala employed over 200 robots. It was streamed on social ‍media platforms RedNote, Sina Weibo, TikTok and its Chinese version Douyin. Chinese-language television networks HTTV and iCiTi TV also broadcast the performance.

"When robots begin to understand Lunar New Year and begin to have a sense of humor, the human-computer interaction may come faster than we think," Ma Hongyun, a photographer and writer with 4.8 million followers on Weibo, said in a post.

Agibot, which says ⁠its humanoid robots are designed for a range of applications, including in education, entertainment and factories, plans to launch an initial public offering in Hong Kong, Reuters has reported.

State-run Securities Times said Agibot had opted out of the CCTV gala in order to focus spending on research and development. The company did not respond to a request for comment.

The company demonstrated two of its robots to Xi during a visit in April last year.

US billionaire Elon Musk, who has pivoted automaker Tesla toward a focus on artificial intelligence and the Optimus humanoid robot, has said the only competitive threat he faces in robotics is from Chinese firms.