Microsoft Outlook Copilot: Privacy Concerns and User Reports Source: Let's Do Science Microsoft Outlook is re-enabling Copilot for some users even after they disable the feature. Multiple community reports show the assistant returns in the new Outlook for Windows, the Outlook Web client, and mobile apps, typically on consumer Microsoft accounts and Microsoft 365 Family subscriptions. Users report toggling the in-app Copilot aktivieren switch … Continue reading Eye on AI: Microsoft Outlook Copilot, Privacy Concerns and User Reports

Advancements in Cancer Care: AI and Chemo Prediction Insights Source: Let's Do Science Technion researchers, led by Gil Shamai, developed a deep-learning system that predicts breast cancer recurrence risk and the probability that a patient will benefit from adjuvant chemotherapy using routine pathology slides. The model estimates the Oncotype DX 21-gene recurrence score from digitized H&E whole-slide images combined with … Continue reading Eye on AI: Advancements in Cancer Care, AI and Chemo Prediction Insights

How AI-Generated Images are Evolving Source: Lifehacker AI-generated images are getting harder and harder to spot. There are the usual tells, of course, but those tells are minimizing with each new AI image model. You could count on AI-generated hands including too many or too few fingers on people, for example, nowadays, that's not necessarily the … Continue reading How AI-Generated Images are Evolving

OpenAI Model Did Better Than ER Doctors at Diagnosing Patients Source: NPR A patient shows up at the hospital with a pulmonary embolism — a blood clot that has traveled to the lungs. After initially improving, their symptoms start to worsen. The medical team suspects the medication isn't working. In steps artificial intelligence — with … Continue reading Eye on AI: OpenAI Model Did Better Than ER Doctors at Diagnosing Patients

Google Expands Pentagon’s Access to its AI After Anthropic’s Refusal Source: Tech Crunch Google has granted the U.S. Department of Defense access to its AI for classified networks, essentially allowing all lawful uses, according to multiple news reports. This deal follows Anthropic’s public stand against the Trump administration after the model maker refused to grant the DoD … Continue reading Eye on AI: Google Expands Pentagon’s Access to its AI After Anthropic’s Refusal

Micro-Interactions Redefined: Meet Clicky AI Assistant Source: Let's Do Science Clicky, an open-source AI "cursor buddy" created by Farza Majeed, runs as a tiny, animated assistant that lives next to your cursor, observes the screen, talks back, and points to UI elements. The experience emphasizes immediate, local context rather than a separate chat window, turning … Continue reading Eye on AI: Micro-Interactions Redefined, Meet Clicky AI Assistant

Claude AI Agent Deletes PocketOS Entire Database in 9 Seconds Source: The Independent An AI agent powered by Anthropic’s leading Claude model has deleted a company’s entire production database, leaving customers unable to access key data. PocketOS, which provides software for car rental businesses, suffered a massive outage over the weekend after the autonomous artificial intelligence tool wiped the database … Continue reading Eye on AI: Claude AI Agent Deletes PocketOS Entire Database in 9 Seconds

ChatGPT Predicts Jobs Likely to Disappear by 2036 Source: Let's Do Science ChatGPT predicted a short list of roles likely to disappear within 10 years, naming Data Entry Clerks, Telemarketers, Cashiers, Travel Agents, and Bank Tellers as primary examples. The piece highlights wage baselines (for example Data Entry Clerks mean salary $40,130) and frames the shift as driven by automation and conversational AI. Technical … Continue reading Eye on AI: ChatGPT Predicts Jobs Likely to Disappear by 2036

Eye on AI: DeepMind Maps Internet-Based Attacks on AI Agents by Let's Do Science Google DeepMind released a paper and accompanying blog that define and demonstrate a new attack class they call "AI Agent Traps", mapping how the open internet can be used to manipulate autonomous agents. The researchers present six distinct trap categories, provide proof-of-concept examples, and … Continue reading Eye on AI: DeepMind Maps Internet-Based Attacks on AI Agents

AI Vendors Avoid Fixing Critical Security Flaws by Let's Do Science AI vendors are deflecting responsibility for security flaws, labeling exploitable behaviors as "working as intended" and choosing documentation changes or token bounties over systemic fixes. Researchers demonstrated hijacks against Claude Code Security Review, Gemini CLI Action, and GitHub Copilot that can exfiltrate API keys and tokens. Vendors paid … Continue reading Eye on AI: AI Vendors Avoid Fixing Critical Security Flaws