&w=3840&q=100)
Microsoft warns govt, firms of cyberattacks on document-sharing software
The FBI on Sunday said it is aware of the attacks and is working closely with its federal and private-sector partners, but offered no other details.
In an alert issued on Saturday, Microsoft said the vulnerabilities apply only to SharePoint servers used within organizations. It said that SharePoint Online in Microsoft 365, which is in the cloud, was not hit by the attacks.
The Washington Post, which first reported the hacks, said unidentified actors in the past few days had exploited a flaw to launch an attack that targeted US and international agencies and businesses.
The hack is known as a "zero day" attack because it targeted a previously unknown vulnerability, the newspaper said, quoting experts. Tens of thousands of servers were at risk.
Microsoft did not immediately respond to a request for comment.
In the alert, Microsoft said that a vulnerability "allows an authorized attacker to perform spoofing over a network." It issued recommendations to stop the attackers from exploiting it.
Microsoft said on Sunday it issued a security update for SharePoint Subscription Edition, which it said customers should apply immediately.
It said it is working on updates to 2016 and 2019 versions of SharePoint. If customers cannot enable recommended malware protection, they should disconnect their servers from the internet until a security update is available, it said.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Mint
an hour ago
- Mint
Mint Explainer: Can AI diagnose you better than a doctor?
Few sectors are witnessing AI's disruptive power as dramatically as healthcare. Trained on millions of historical patient records, advanced AI models are now matching—and in some cases surpassing—human expertise. With real-time analysis across medical histories, imaging, and genetic profiles, AI transforms clinical complexity into diagnostic clarity. AI spots patterns even seasoned physicians might overlook. Earlier this month, Microsoft unveiled a platform–Microsoft AI Diagnostic Orchestrator (MAI-DxO)—that can diagnose medical cases with an accuracy rate of up to 85.5%, far exceeding the 20% accuracy of experienced physicians under the same conditions. Mint explains the role and impact of AI in healthcare and diagnostics. Does Microsoft's MAI-DxO improve diagnostic accuracy? MAI-DxO leverages vast datasets, advanced probabilistic reasoning, and continuous learning from clinical cases to surpass human diagnostic accuracy. Unlike doctors who rely on experience and pattern recognition, MAI-DxO evaluates multiple variables simultaneously—lab results, symptoms, imaging—and predicts outcomes. MAI‑DxO was tested on 304 real-world clinical scenarios sourced from the New England Journal of Medicine. In these cases, MAI‑DxO achieved an accuracy rate of 85.5%, surpassing the average 20% success rate of 21 experienced physicians from the UK and the US. Instead of relying on a single model, MAI-DxO coordinates multiple large language models (LLMs) that interact like a team of doctors, reviewing, challenging, and refining each other's suggestions before settling on a final diagnosis. Microsoft has emphasised that the tool is not designed to replace doctors, but to work alongside them. What medical datasets does MAI-DxO use? MAI-DxO is trained on anonymised patient records, peer-reviewed literature, medical guidelines and global clinical datasets from hospitals and research institutions. Its database spans symptoms, biomarkers, comorbidities, and disease trajectories across demographics, allowing it to contextualize findings in real-world scenarios. So if a patient suddenly feels numbness on her left side, it could indicate stroke, nerve damage, heart problem or some other issue. Doctors will wait for test results (blood and scans) to diagnose the problem, while the AI platform will interpret it faster, helping the doctor make quicker decisions. MAI-DXO's strength lies in its ability to absorb medical nuance across specialties—mirroring the cumulative expertise of several doctors and researchers. Will AI replace doctors? AI will not replace but complement doctors and other health professionals. They need to navigate ambiguity and build trust with patients. Clinical roles will evolve with AI, giving medical staff the ability to automate routine tasks, identify diseases earlier, personalize treatment plans, and potentially prevent some diseases altogether. For consumers, they will provide better tools for self-management and shared decision-making. These will be particularly helpful in remote areas where there's a shortage of doctors. MAI-DxO operates on statistical inference rather than intuition. It learns to recognise patterns and contextual subtleties from data, simulating judgment. While human intuition draws on tacit knowledge—emotion, empathy, gut feel—AI taps structured data and outcome probabilities. Platforms like MAI-DxO can augment decision-making. These are especially useful when doctors second-guess themselves—making MAI-DxO an ally, not a rival. What risks do AI-driven diagnostics pose for patient care? Key risks include overreliance, bias from skewed training data, and lack of transparency in AI reasoning. If doctors defer judgment too easily, subtle clinical clues could be overlooked. Additionally, AI platforms recommendations may vary based on demographics or comorbidities not well represented in its training data, potentially exacerbating health disparities. For instance, if they are trained only on American or European patient data, they might overlook or misinterpret common conditions among patients in South Asia, say tuberculosis (TB) or type 2 diabetes or complications caused by chewing tobacco. There's also the risk of false positives and patient anxiety. Most critically, AI tools must be accountable — patients need clarity on how diagnoses are made. The challenge lies in integrating AI without ignoring the human touch in care. What are the regulatory frameworks for AI deployment in hospitals? In India, every AI tool must be approved by the Delhi-based Central Drugs Standard Control Organisation (CDSCO)— equivalent to the US Food and Drug Administration (FDA). Globally, regulatory bodies like the FDA (US), European Medicines Agency (EMA-Europe) are developing guidelines for AI in medicine. Back home, AI platforms need clearance as a software as a medical device (SaMD), requiring evidence of clinical safety, efficacy, and data privacy compliance. What is the role of AI in the future of medicine? Current developments position them as early warning systems, helping doctors rather than making final decisions. According to the World Economic Forum 4.5 billion people are currently without access to essential healthcare services and a health worker shortage of 11 million is expected by 2030. AI has the potential to help bridge that gap. AI can also assist paramedics in situations where a patient is being transferred via an ambulance. AI models trained on factors such as a patient's mobility, pulse and blood oxygen levels, chest pain, etc., can relay information to doctors, helping them make decisions faster. AI can detect early signs in an individual that are likely to result in diseases like Alzheimer's, heart disease, kidney disease, and so on. This could help doctors suggest preventive action. Can MAI-DxO evolve into a real-time medical assistant across specialties? At present, MAI-DxO is a demonstration of AI capability and research. With real-time access to patient records, labs, and imaging, MAI-DxO could become an assistant, offering diagnostic suggestions, tracking progress, and alerting doctors to anomalies as they emerge. Integrated into electronic health systems, it could support rounds, triage, and even remote consultations. Much like AI embedded medical devices (ultrasound devices, bedside X-ray machines) are now being used in some hospitals, including Max healthcare, to interpret scans and assist doctors. The key lies in continual updates, feedback loops and building trust.
India.com
an hour ago
- India.com
This IT giant is buying human waste for Rs 140000000000 due to..., name is..., it is owned by...
As part of an expansive sustainability initiative, Microsoft is acquiring 4.9 million metric tonnes of organic waste, which includes human waste, manure, food scraps, paper sludge and agricultural residue. As per a report from Inc., the deal is with climate-tech startup Vaulted Deep, or what the company calls 'putting waste to work.' Microsoft hopes to significantly reduce its global carbon footprint and move closer to its grand environmental goals through this commitment. Microsoft's partnership comes at a time while the tech company urgently seeks new creative options to address the massive energy usage of its growing AI operations, and to cut the carbon emissions tied to them. The company produced 75.5 million tons of CO₂ equivalent from 2020 to 2024, much of which is attributed to its recent rise in AI infrastructure load. At a vague market value of $350 per ton for carbon removal, this constitutes one of the largest investments in waste-to-carbon technology, according to the Wall Street Journal. According to the official website of Vaulted Deep, Vaulted is a dual waste management solution: we remove carbon from the atmosphere while protecting nearby towns from contamination.
Indian Express
an hour ago
- Indian Express
Google CEO Sundar Pichai on AI talent war: ‘Our talent retention is healthy'
Google CEO Sundar Pichai says the company continues to attract top AI talent, brushing off concerns about the ongoing AI talent war in Silicon Valley, where top tech companies are poaching talent from each other to stay ahead in the intensifying AI race. During Alphabet's second-quarter earnings call, Pichai said Google has seen such moments before and that its core metrics remain 'healthy.' 'We continue to look at both our retention metrics, as well as the new talent coming in, and both are healthy,' Pichai said on the earnings call. 'I do know individual cases can make headlines, but when we look at the numbers deeply, I think we are doing well through this moment.' Pichai also emphasised that Google understands what it takes to keep top AI researchers satisfied and it isn't all about the money. He cited the company's investments in greater access to compute, referring to the latest and most advanced computer chips. At a time when the world's leading tech companies are laying off parts of their workforce due to the rise of artificial intelligence, they are simultaneously hiring specialised AI researchers, both from top PhD programs in machine learning and from among industry veterans who have created platforms at leading tech companies. From Meta and OpenAI to Google and Microsoft, these companies are aggressively poaching top AI talent from each other. With Meta launching its Superintelligence Lab and hiring Apple's Foundation Models lead Ruoming Pang with a reported package exceeding $200 million, talent and compute have become the defining currencies in the AI world. In recent months, Meta also recruited Scale AI founder Alexandre Wang and a dozen veterans from OpenAI and DeepMind, signaling its deepening investment in AI. In June, Adam Sadovsky. who spent nearly 18 years at Google, most recently as a distinguished software engineer and senior director at DeepMind left the company to join Microsoft. He now serves as a corporate vice president at Microsoft AI. This current AI hiring frenzy echoes the intense 'war for talent' between Microsoft and Google in the 2000s, triggered by Google's meteoric rise. At the time, Google's focus on search attracted some of the brightest engineering minds, forcing Microsoft to revise its hiring and retention strategies. In 2005, Microsoft even sued Google after executive Kai-Fu Lee left the company to join Google, a milestone case that highlighted the fierce competition between the two tech giants over top talent.
