Breaking AI & Tech News Analyzed

China has emerged as a leader in global AI governance, contrasting with the US, which is pursuing AI development in a 'wild west' manner, according to Prof Dame Wendy Hall, a former UN and UK government adviser. Hall told the House of Commons business and trade committee that China is backing multinational attempts to introduce global governance of AI, while the US has set up a race between profit-hungry companies that rely on hype.Hall, who is director of the Web Science Institute at the University of Southampton, said Chinese AI researchers are efficient, innovative, and willing to release their models on an open-source basis. However, she noted that it has become increasingly difficult for UK experts to collaborate with China on research, limiting her academic freedom.The UK's reliance on US tech companies, including Google, Microsoft, OpenAI, and Amazon, risks a repeat of the Post Office Horizon scandal, warned Neil Lawrence, Cambridge University's DeepMind professor of machine learning. He expressed concerns that the UK is outsourcing AI model development to private billionaires with zero loyalty to the British state and consumer.Hall and Lawrence also highlighted that promises from US-backed tech companies may not be delivered as planned. For example, OpenAI has put a UK datacentre project on hold, and a government plan to open a large UK sovereign AI datacentre is behind schedule.The tech industry has identified a lack of power as a key problem, with Microsoft saying a planned datacentre in the north of England will not come online until at least 2033 due to a shortage of power from the grid.

The Evolution of Enterprise AutonomyMicrosoft is quietly pivoting from reactive AI assistants to proactive, autonomous agents within its ecosystem. The tech giant is currently testing a new feature set for Microsoft 365 Copilot that mimics the capabilities of the open-source OpenClaw agent. This move signals a strategic shift toward "always-on" intelligence that can execute multistep tasks autonomously, rather than merely responding to user prompts. Microsoft's "Always-On" Copilot StrategyThe core innovation of this potential new agent is its ability to function continuously. Unlike previous iterations that required active user engagement, this tool would be designed to take actions at any time, effectively acting as a persistent digital assistant. Microsoft has confirmed to The Information that the focus is on enterprise customers, specifically addressing the security concerns that have historically plagued open-source alternatives. Autonomous Execution: Capable of handling multistep workflows without constant supervision. Enterprise Focus: Prioritizing security controls over the flexibility of open-source tools. Integration: Built directly into the existing Microsoft 365 ecosystem. Cloud vs. Local: The Hardware ImplicationWhile the source material suggests a comparison with OpenClaw—which runs locally on hardware like the Mac Mini—Microsoft has not confirmed if this new agent will be local or cloud-based. However, the trend is clear. The company previously launched Copilot Cowork (powered by Anthropic's Claude) and Copilot Tasks, both of which operate in the cloud. The potential shift to a local execution model would explain the recent surge in Mac Mini sales, as users seek hardware capable of running these resource-intensive, privacy-focused agents. Why This Matters for Enterprise SecurityThe primary driver for this development is the "trust gap" in enterprise AI. Open-source agents like OpenClaw offer powerful automation but carry significant security risks. By creating a proprietary version, Microsoft aims to offer the autonomy of open-source tools with the governance of a major corporation. This aligns with Microsoft's broader strategy of anchoring AI experiences in security, governance, and trust, reducing the friction of daily operations for enterprise workers. Expectations for Microsoft Build 2026Industry analysts predict that this new agent—or an upgraded version of existing tools—will be a centerpiece of the upcoming Microsoft Build conference in June. While the company remains tight-lipped about the specifics, the spokesperson's confirmation that they are "experimenting" with broader orchestration and autonomy suggests a major reveal is imminent. This development could redefine how businesses interact with their software stack, moving from a tool-based model to an agent-based model.

Anthropic, a San Francisco-based AI startup, has announced that its latest AI model, Claude Mythos, has proven highly effective in exposing software weaknesses. The model has identified thousands of vulnerabilities in commonly used applications, many of which have no patch or fix available.According to Mike Krieger of Anthropic Labs, the company has decided not to release Mythos to the public due to concerns about its potential misuse by hackers. Instead, Anthropic is collaborating with cybersecurity specialists and engineers in the open-source community to utilize the model as a defensive tool.The oldest vulnerability uncovered by Mythos dates back 27 years, and none were previously noticed by their creators. Anthropic has shared a version of Mythos with cybersecurity companies CrowdStrike and Palo Alto Networks, as well as with Amazon, Apple, and Microsoft, in a project dubbed “Glasswing”.“AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities,” Anthropic said in a blog post. “The fallout – for economies, public safety, and national security – could be severe.”The project aims to protect critical infrastructure from cyber threats, with approximately 40 organizations involved in the design, maintenance, or operation of computer systems. Anthropic is providing about $100m worth of computing resources for the mission.

The Accelerator's Withdrawal: A Signal of Loss of ConfidenceDelve's relationship with Y Combinator has officially ended following a series of damaging allegations regarding compliance and data security. This severance marks a significant blow to the startup's credibility, compounded by the distancing actions of other major investors like Insight Partners.The Catalyst: Anonymous Allegations and Data BreachesThe controversy stems from an anonymous Substack campaign by "DeepDelver," which accused the company of misleading clients about regulatory compliance and passing off open-source tools as proprietary technology. These claims were further fueled by a security researcher's ability to access sensitive Delve data and a malware incident involving a customer, LiteLLM.YC's Response: Delve was removed from the accelerator's portfolio directory, with COO Selin Kocalar confirming the split on X.Insight Partners: The firm initially deleted posts about its investment but later restored the primary blog entry.The Defense: A Coordinated Attack or Operational Failure?In a bid to set the record straight, Delve's leadership team, including CEO Karun Kaushik, claims the attacks are a coordinated smear campaign orchestrated by an attacker who exfiltrated internal data. They argue that the "evidence points to a malicious attack rather than a genuine whistleblower."However, the company also acknowledged "growing too fast and falling short of our own standard." To mitigate the damage, Delve has hired a cybersecurity firm, offered complimentary re-audits to customers, and clarified that their open-source usage is compliant with Apache 2.0 licensing.Future Outlook: Rebuilding Trust in a Fragile EcosystemThe departure from Y Combinator suggests that the startup's growth trajectory is now in jeopardy. For a compliance-focused company, trust is the primary currency; the current allegations threaten to devalue this currency permanently. The coming months will determine if Delve can survive this reputational crisis or if it will become a cautionary tale in the compliance tech sector.