🔥 The Week in AI News - March 16 - 20, 2026

DoorDash, Higgsfield, NVDIA, Google AI, Google Stitch, Adobe Firefly, OpenAI, Canva, Meta, Fitbit, Microsoft, Rebel Podcasting and more.

▪️ DoorDash Turned Couriers Into AI Data Collectors With a New Tasks App

DoorDash launched a side app that paid drivers to capture real-world video, quietly building a human-powered pipeline for training AI models.

▪️ Higgsfield Launched ‘Arena Zero,’ an AI-Native Streaming Platform

Higgsfield introduced a new kind of streaming platform designed entirely around AI-generated content and interactive media experiences.

▪️ NVIDIA Introduced NemoCLAW for Enterprise AI Control

NVIDIA unveiled NemoCLAW, a framework aimed at giving companies tighter control, safety, and customization over large language models.

▪️ Google AI Teased Its Latest Model Advancements on X

Google shared updates hinting at continued rapid iteration across its AI model stack, signaling more competition at the frontier level.

▪️ Google Labs Released Stitch for AI UI Design

Google launched Stitch, a tool that turned prompts into full interface designs, pushing AI deeper into product and frontend workflows.

▪️ Adobe Brought Custom Firefly Models to Public Beta

Adobe opened access to custom-trained Firefly models, letting creators generate visuals aligned with their own brand styles.

▪️ OpenAI Tested an Adult Mode for ChatGPT Writing

OpenAI explored a more permissive mode for generating explicit content, raising new questions around boundaries and monetization.

▪️ Meta Struggled With Rogue AI Agents Inside Its Systems

Meta reportedly faced issues controlling autonomous AI agents, highlighting how quickly agent behavior can drift off-script.

▪️ Gamma Added AI Image Generation to Compete With Canva

Gamma expanded into AI visuals, aiming to become a full-stack creative tool instead of just a presentation platform.

▪️ Fitbit Introduced an AI Health Coach That Reads Medical Data

Fitbit rolled out a coaching system that analyzed medical records, moving wearables closer to proactive health management.

▪️ Perplexity Released ‘Comet’ for iOS

Perplexity launched a new mobile experience designed to make AI search feel faster, more conversational, and always-on.

▪️ Google Expanded Its Personal Intelligence Feature Across the U.S.

Google scaled its personal AI layer to more users, deepening its push toward a unified assistant across devices and services.

▪️ Merriam-Webster and Britannica Sued OpenAI Over Training Data

Major reference publishers filed a lawsuit claiming unauthorized use of their content in AI model training.

▪️ Microsoft Launched MAI-Image-2 as a Top-Tier Image Generator

Microsoft introduced a new image model positioned among the best, intensifying the race for photorealistic AI visuals.

▪️ NVIDIA Built a Secure Alternative to OpenCLAW

NVIDIA developed its own secure framework to address vulnerabilities in AI deployment, focusing on enterprise-grade protection.

▪️ Memories.ai Built a Visual Memory Layer for Devices

A startup worked on giving AI persistent visual memory, aiming to power wearables and robots with contextual awareness.

▪️ NVIDIA’s Networking Division Quietly Became a Multi-Billion Dollar Force

NVIDIA’s networking arm scaled rapidly, positioning itself as a second powerhouse alongside its chip business.

▪️ World Launched a Tool to Verify Humans Behind AI Agents

A new system aimed to confirm real human identity in transactions, responding to the rise of autonomous shopping agents.

▪️ Rebel Audio Debuted an AI Podcasting Tool for Beginners

Rebel Audio introduced a simplified platform that let first-time creators generate and publish podcasts with AI assistance.

▪️ The Department of Defense Rejected Anthropic Over Risk Concerns

The Pentagon flagged Anthropic’s safety constraints as a potential liability, complicating its path into defense contracts.

▪️ The Pentagon Explored Alternatives to Anthropic’s AI

U.S. defense agencies began evaluating other AI providers, signaling a shift in who might power national security systems.

▪️ Cloudflare Predicted Bots Would Outnumber Humans Online by 2027

Cloudflare’s CEO warned that AI-driven traffic would soon dominate the internet, reshaping how platforms measure reality.

▪️ Jeff Bezos Planned a $100B AI Roll-Up of Manufacturing Firms

Bezos reportedly explored using AI to modernize legacy industries at scale, targeting inefficiencies across old-school manufacturing.