AI Video
AI video agents and video generation APIs solve different problems in synthetic media pipelines. Here's how they differ architecturally, when to use each, and what it means for building scalable AI video workflows.
AI Video
AI video agents that generate content without grounding produce plausible but inaccurate output. A research layer powered by the Model Context Protocol (MCP) gives these agents access to verified data before rendering, improving factual reliability in synthetic media.
AI Agents
The UK's AI Security Institute, working with OpenAI and Anthropic, found AI agents can autonomously fabricate fake online identities and attempt hacking — raising urgent questions about synthetic personas and digital trust.
AI Governance
New research finds half of Australian and New Zealand organisations deploy AI agents without governance frameworks, just as deepfake-driven fraud and impersonation threats intensify across the region.
AI Agents
A Harvard and Perplexity study finds AI agents perform 26 minutes of autonomous work per session versus 33 seconds for traditional search, signaling a fundamental shift in how users delegate complex tasks to AI systems.
Microsoft Research
Microsoft Research unveils Webwright, a terminal-native web agent framework that nearly doubles GPT-5.4's baseline score on the Odysseys benchmark, jumping from 33.5% to 60.1%.
Microsoft
Microsoft's new Fara1.5 family of browser computer-use agents (4B/9B/27B) outperforms OpenAI Operator and Gemini 2.5 Computer Use on the Online-Mind2Web benchmark, marking a major step for open-weight web agents.
Alibaba
Alibaba's new Zhenwu M890 chip roadmap is designed specifically around AI agent workloads, signaling a shift from raw training power to inference-heavy, agentic compute that could reshape the AI hardware race.
AI Agents
Stripe's Link digital wallet now supports autonomous AI agents, letting them complete purchases on behalf of users — a major step toward agentic commerce infrastructure.
OpenAI
OpenAI's GPT-5.5 is a fully retrained agentic model scoring 82.7% on Terminal-Bench 2.0 and 84.9% on GDPval, signaling a major step forward in autonomous coding and real-world task execution capabilities.
Nous Research
Nous Research's Hermes model introduces adaptive learning capabilities that let AI agents personalize to individual users, marking a shift from static assistants toward genuinely customizable AI systems.
Nvidia
NVIDIA's Nemotron Personas dataset on Hugging Face offers demographically-grounded synthetic personas for building culturally-aware Korean AI agents, tackling the challenge of localized persona simulation with structured synthetic data.