Full Summary
This Sunday morning, a major shift in AI development is underway, with companies like OpenAI and Anthropic now battling over the *cost* of digital labor, not just raw power. Both Il Sole 24 ORE and Laodong.vn highlight this strategic pivot, moving towards platforms and business services aimed at replacing human tasks at the lowest possible cost. Glovo has launched a new AI shopping assistant, confirmed by The Independent Uganda, that integrates its marketplace directly into ChatGPT and Claude. This allows users to find products, compare prices, and place orders through natural conversation, like asking for "a last-minute gift for a coffee lover under €50." Once connected, the AI remembers preferences and presents up to five personalized recommendations. But then, MakeUseOf reveals a significant challenge: AI's memory features can make it difficult to erase false information. An experiment successfully taught ChatGPT, Claude, and Gemini the same lie about a favorite author, demonstrating how easily misinformation can embed itself into their persistent memory. What nobody expected is Microsoft's aggressive move, with CEO Satya Nadella confirming plans to transform Copilot into a single "super app" AI by the end of 2026, merging its various individual versions. Laodong.vn reports this aims to create a comprehensive background task platform for both general users and businesses, potentially coming pre-installed with Windows 11 updates. This is a direct play to become the "default AI interface." Finally, Tech Times reports that California's AI Transparency Act is now in effect, requiring AI systems with over one million monthly users to embed machine-readable data and offer public detection tools. Midjourney, a popular AI image generator, currently lacks these C2PA content credentials, facing potential civil penalties of $5,000 per violation, per day. This means that while AI is becoming more integrated into your daily shopping and work, you'll also have to navigate its persistent memory issues and soon, legally mandated disclosures on AI-generated content.