Perplexity Launches Computer Control as Governments Race Toward AI Adoption

In partnership with

Perplexity Unveils “Perplexity Computer”

Perplexity introduced Perplexity Computer, a new agentic system designed to move beyond answering questions and toward executing tasks across a live desktop environment. Instead of returning links or summaries, the system can browse, click, extract data, and complete multi-step workflows — effectively acting as an autonomous research and operations assistant.

The shift signals Perplexity’s evolution from AI search engine to action-oriented agent platform, where the value isn’t just better answers, but task completion. As more AI products add “computer use” capabilities, the battleground is moving from intelligence to execution reliability and real-world control.

Summer Yue Flags Alignment Risks After Agent Deletes Email Archive

Researchers at Meta shared a cautionary case study while testing advanced agent systems (including open computer-use frameworks like OpenClaw): during a task execution sequence, an autonomous agent mistakenly deleted an entire email archive after misinterpreting instructions.

The incident underscores a critical alignment challenge — once agents gain real system permissions (file access, inbox control, tool execution), small reasoning errors can escalate into irreversible actions. Meta emphasized that stronger guardrails, sandboxing, permission scoping, and reversible action layers are essential before deploying high-agency systems in production environments.

As agent capabilities expand from “suggest” to “act,” safety is shifting from model alignment alone to operational containment, auditability, and fail-safe design.

UK Treasury Enlists AI Expertise to Transform Public Services

The UK Treasury has officially partnered with the Tony Blair Institute for Global Change and major tech firms like IBM, Faculty AI (part of Accenture), and others to guide AI adoption in government services. This strategic move aims to make the UK one of the fastest adopters of AI among the G7, embedding AI across sectors such as health, defence, policing, and the courts to improve efficiency and decision-making. However, the initiative has drawn criticism from civil liberties groups concerned about private tech involvement in public policy decisions. Leaders emphasize that scalable AI solutions must balance innovation with ethical and transparent deployment, especially when the tech touches citizens’ daily lives.

AI Planning Agent Helps Tasmanian Residents Navigate Urban Growth

The Huon Valley Council in Tasmania launched an AI-powered planning tool designed to answer residents’ questions about local planning rules, building permits, and procedural requirements. Available 24/7 via phone, this AI tool gives community members accurate, timely guidance without requiring them to visit council offices or wait in long phone queues. Beyond helping residents file applications faster, this system also reduces staff workload, enabling human officials to focus on complex cases. The success in Huon Valley is inspiring other councils in Tasmania and across Australia to introduce similar solutions that make local governance more responsive and people-friendly.

AI Adoption Soars Across U.S. Federal Agencies, Reshaping Public Services

A recent analysis shows that AI use across U.S. federal government agencies has nearly doubled in the past few years, expanding from isolated pilots into mission-critical functions. From the Department of Veterans Affairs to Homeland Security, agencies are deploying AI to streamline internal operations, improve customer service, and enhance accuracy in decision-making. Policy reforms and guidance from the White House and Office of Management and Budget have helped standardize AI procurement and governance, bringing more accountability and structure to these technologies. For everyday citizens, this means faster responses to requests, improved public services, and more transparent government workflows when interacting online or in person.

What to Watch this Week

 🎙️AI on Fire: Real Builders. Real Heat.

Stories from people building AI Agents. Explore all episodes here

🏆 The Leaderboard Never Sleeps

The global ranking of AI agents is shifting every day. Who’s on top? Who just dropped?

👉 Explore the Leaderboard →

🗺️ The Map of AI Agents (Live & Growing)

We’re charting the entire AI agent ecosystem — thousands of options across categories.
Your next agent is already on the map.

👉 Explore the Landscape Map →