Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Here’s your AI briefing for Tuesday, March 24, 2026. Seven stories from the past week that matter if you work with, build on, or compete against AI. No filler.
The thread running through this week’s news is consistent: AI stopped being a tool you point at problems and became something that runs things on its own. It’s publishing websites, replacing support teams, moderating billions of posts, and in one case, putting a dead actor back on screen. Here’s what happened.
WordPress.com gives AI agents full control over your website
WordPress.com, the platform behind over 43% of all websites, opened the door to fully autonomous AI publishing last Friday. AI agents can now draft, edit, and publish posts, manage comments, update metadata, and reorganize content using natural language commands.
The update builds on MCP (Model Context Protocol) support introduced last autumn. AI assistants previously had read access to WordPress.com sites. Now they have full write access, meaning a site can be created and run almost entirely by an AI agent with a human at the top level.
WordPress.com sees 409 million unique visitors per month. If even a fraction of those sites shift to AI-managed content, the character of the web changes fast. The question isn’t whether this will happen but how quickly, and whether readers will know the difference.
Source: TechCrunch
Meta launches AI support assistant and phases out human moderators
Meta rolled out its AI Support Assistant across Facebook and Instagram this week, offering 24/7 automated help for account issues and signalling that human content moderation contractors are being scaled back.
The assistant handles scam reports, impersonation complaints, content appeals, privacy settings, password resets, and profile updates. Meta says it typically responds in under five seconds.
The automation extends to content enforcement too. Meta says the AI catches severe violations like scams faster and more accurately, with fewer over-enforcement mistakes. Whether that claim holds up under scrutiny is another question. What’s clear is that AI is replacing humans on Meta’s moderation stack, and the contractors who handled this work are being phased out.
Source: Meta Newsroom
Signal’s creator is bringing end-to-end encryption to Meta AI
Moxie Marlinspike, the engineer who brought end-to-end encryption to WhatsApp, is doing it again for AI. His startup Confer is integrating its privacy technology into Meta AI so that conversations remain encrypted and inaccessible even to Meta itself.
Marlinspike’s argument is blunt: AI chat apps have become some of the largest centralized data lakes in history, containing medical records, financial details, and unfiltered private thoughts. None of it is currently private. The AI company can read it, their employees can read it, and a subpoena can surface it.
Ten years ago, the Signal Protocol reached billions through WhatsApp. Now Confer’s approach could do the same for AI, making privacy-preserving chat a default rather than a niche feature. Confer continues to operate independently while this integration is built out.
Source: Confer Blog
Microsoft’s MAI-Image-2 hits the top 3 for text-to-image generation
Microsoft launched MAI-Image-2 this week, its second-generation text-to-image model. It’s already ranked #3 globally on the Arena.ai leaderboard, behind only a small group of specialist image labs.
The model was built with photographers and designers in mind. Improvements include better photorealism with natural lighting and accurate skin tones, reliable in-image text generation, and richer scene composition for complex or surreal concepts. Microsoft says creatives will spend less time fixing outputs in post-production.
MAI-Image-2 is live in the MAI Playground and rolling out to Copilot and Bing Image Creator. API access is available for enterprise customers including WPP, with broader developer access coming.
Source: Microsoft AI
AI has fuelled $70 billion in data centre M&A this year
The AI infrastructure buildout is producing one of the biggest M&A waves the data centre industry has seen. Bloomberg Law reports $70 billion worth of deal talks have already taken place in 2026, driven by SoftBank, DigitalBridge, and others racing to lock in compute capacity.
The logic is straightforward: training and running frontier AI models requires enormous, reliable power and bandwidth, and demand outpaces what most operators can build on their own. Buying existing capacity is faster. The result is a consolidation wave that mirrors what happened to cloud infrastructure in the 2010s, only faster.
For context, the entire global data centre M&A market in 2023 was roughly $60 billion for the full year. The 2026 pace already exceeds that in one quarter.
Source: Bloomberg Law
Amazon brings Alexa Plus to the UK
Amazon’s AI-upgraded Alexa Plus landed in the UK on March 19 as part of an Early Access program, the first European launch for what Amazon calls a complete rethink of its voice assistant.
Alexa Plus is conversational, remembers context across sessions, handles complex multi-step tasks, and works across devices without requiring a wake word between every sentence. During Early Access it’s free. After that it’s free for Prime members or £19.99/month otherwise.
UK customers have used Alexa over 114 billion times in the last three years. Amazon is betting that upgrading that installed base with a generative AI brain gives it a distribution advantage that pure AI startups can’t match.
Source: Amazon Newsroom
Val Kilmer is starring in a film, AI-generated, with his family’s blessing
Val Kilmer, who died in 2025 from throat cancer, will play a leading role in a new film thanks to generative AI and the approval of his family. The film, As Deep as the Grave, cast Kilmer as Father Fintan years before his death. He was never well enough to shoot a single scene.
Director Coerte Voorhees reconstructed the performance using younger images provided by Kilmer’s family, footage from his later years, and a reconstruction of his voice, which was damaged by cancer treatment. His daughter Mercedes and son Jack both support the project. “He really thought it was an important story that he wanted his name on,” Voorhees said.
Kilmer will appear in a “significant part” of the finished film alongside Tom Felton, Abigail Lawrie, Wes Studi, and Abigail Breslin. This is one of the first AI-generated posthumous performances with explicit estate approval, and it raises real questions about where the lines for digital resurrection should be drawn.
Source: Variety
Frequently asked questions
Can AI agents now automatically publish blog posts on WordPress?
Yes. WordPress.com now allows AI agents to draft, edit, and publish content via the Model Context Protocol (MCP). Site owners can direct agents using natural language commands, giving AI systems full write access to their websites.
Is Meta replacing all human content moderators with AI?
Meta is scaling back human moderation contractors as its AI Support Assistant takes on more work. The company says the AI catches violations faster with fewer errors, though independent auditing of that claim is limited.
Will Meta AI conversations become end-to-end encrypted?
Moxie Marlinspike’s startup Confer is working to integrate end-to-end encryption into Meta AI. This is still in development, but Marlinspike successfully did the same for WhatsApp, so there’s real precedent.
What is Microsoft MAI-Image-2?
MAI-Image-2 is Microsoft’s second-generation text-to-image AI model. It launched in March 2026 and ranked #3 on the Arena.ai global leaderboard. It’s rolling out to Copilot and Bing Image Creator.
Is the AI-generated Val Kilmer film legal?
Yes. The production has explicit approval from Kilmer’s estate and his children. Whether broader AI posthumous performances should require legal frameworks beyond estate consent is a live debate.
Stay in the loop
Every weekday, FridayAIClub.com publishes the AI stories that actually matter. Bookmark us, share this with someone who needs to keep up, or just come back tomorrow. We’ll be here.
