All Updates
Page 517 of 752
March 6, 2026
Copilot Adds Built-in Browser Pane
Microsoft is updating Copilot on Windows with a built-in side-pane browser. Users can open links, research across tabs, and summarize information without leaving the app. This eliminates the need for browser switching during AI-assisted tasks.
69% Africa Biometric Fraud Now AI-Generated
A report states 69% of biometric fraud in Africa's fintech is AI-generated. Fraudsters hijack verified accounts instead of creating fakes. This reveals blind spots in systems focused on signup risks.
AI Powers US Military Targeting Shift
Palantir's Maven with Anthropic Claude enables real-time targeting, compressing weeks to hours in strikes. Defense AI startups like Anduril attract $30B VC surge. Silicon Valley pivots from consumer to military AI amid global conflicts.
Cloudflare One Secures Endpoint to AI Prompts
Cloudflare One delivers unified data security from endpoints to AI prompts. Key additions include RDP clipboard controls, operation-mapped logs, on-device DLP, and API CASB scanning for Microsoft 365 Copilot.
City Detect Raises $13M Series A for AI Urban Safety
City Detect, an AI startup helping cities stay safe and clean, has raised $13M in Series A funding. The company assists local governments in preventing urban decay and is already active in at least 17 cities, including Dallas and Miami.
OpenClaw AI Agents Explode in China
OpenClaw, an open-source AI agent framework, has ignited a 'raising shrimp' trend from Silicon Valley to China, with users aged 2-70 building apps and automating tasks. Tencent offers free installs, while Baidu, Alibaba launch clones like LobsterAI. Interviews reveal uses from kids' mini-programs to quant trading, despite costs and glitches.
US Tightens AI Chip Export Controls
The Trump administration plans new GPU export restrictions to control buyers and drive AI investment back to the US. Nvidia and AMD may require government approval for foreign sales. This aims to bolster domestic AI development.
AI Browsers Vulnerable to Prompt Injection
Perplexity's Comet and OpenAI's ChatGPT Atlas face severe prompt injection vulnerabilities allowing data theft via calendar invites or fake URLs without malware. Zenity Labs' PleaseFix exploits AI's inability to distinguish malicious external prompts. Companies deploy defenses like logout modes but warn of persistent risks in cat-and-mouse security game.
Zhejiang, Guangdong Race for Xi's AI Innovation
China's Zhejiang and Guangdong provinces are competing to lead in AI and core technology breakthroughs, responding to President Xi Jinping's call. This aligns with priorities for the 15th five-year plan outlined during the 'two sessions'. Zhejiang has unveiled an ambitious plan to become a major AI development hub.
AI Boosts Malvertising as Top Malware Vector
Malvertising has surpassed email and direct hacks as the leading global malware delivery vector, per The Media Trust's report. AI is accelerating threats by enabling adaptive malware that adjusts to location, browser, or device. Future risks include AI-assisted evasion tactics and adtech exploitation for payloads.
UK Warns AI Copyright Weakening Harms Creatives
House of Lords committee urges stronger AI copyright laws to protect Britain's Β£124B creative sector. Weakening rules risks trading industries for uncertain tech growth promises. Ministers must prioritize protections.
Claude AI Debuts in US Military Kill Chain
Anthropic's Claude integrated with Palantir AIP used in US-Israel strikes on Iran and Venezuela for intel analysis and decision support. Pentagon signs $200M deal; first commercial LLM confirmed in combat. Highlights accelerating tech-to-battlefield timeline.
WhatsApp Opens Rival AI Chatbots in Brazil
Meta is allowing rival AI companies to offer chatbots on WhatsApp to Brazilian users for a fee. This expansion follows a similar rollout confirmed for Europe just one day prior. It enables third-party AI integration into WhatsApp's massive Brazilian audience.
Anthropic Launches AI Job Loss Warning Tool
Anthropic has launched a new tool to monitor AI's effects on the workforce and warn when jobs might be displaced by AI. The company emphasizes that AI remains far from its theoretical capabilities. It offers insights for policymakers to protect future workers.
China Boosts Tech Funding Policies
China announced measures to support its tech sector, including ChiNext reforms for easier fundraising and refinancing. Officials aim to foster a flexible ecosystem and increase demand for hi-tech products. This addresses high risks and long cycles in technological innovation.
Lawmakers Target Huawei US Loopholes
US lawmakers call on the Trump administration to eliminate loopholes letting Huawei's blacklisted unit operate freely in the US. Huawei Technologies is a Chinese tech giant facing restrictions. This aims to tighten control over its American activities.
Cecilia Liu on Docker's AI MCP Strategy
Cecilia Liu, Senior Product Manager at Docker, discusses leading the MCP strategy to enable secure, scalable AI tooling. In an International Womenβs Day interview, she shares insights on shaping product direction and driving AI innovation. Her role defines how Docker supports AI development workflows.
AI's Limited Lift for SaaS Business
Opinion piece argues current AI adds little real value to enterprise SaaS despite hype, citing failed cases like Builder.ai and low revenue from Salesforce Agentforce. Recommends practical 'left-right' integration where AI augments specific SaaS functions like customer success agents.
Microsoft Pulls Copilot Real Talk Mode
Microsoft discontinued the 'Real Talk' mode in Copilot AI chats weeks after global rollout. It offered the most human-like experience by pushing back, remembering users, and avoiding passive agreement. The plug was pulled despite its engaging personality.
AI Video Gen Faces 6-Hour Queues
AI anime drama industry crippled by compute shortages; generation takes 6 hours, firms resort to night shifts or pricey private APIs costing 200 RMB/min. Examples include Seedance 2.0 overload post-open access, forcing switches to Jimeng 3.5.