AI Agents

AI Safety and Security

Anthropic Warns Self-Replicating AI Could Spread to Your Smart Fridge

Oct 14, 2026

•

18 min read

Anthropic Warns Self-Replicating AI Could Spread to Your Smart Fridge

Your smart refrigerator could one day do more than keep your food cold—it could unknowingly become part of a spreading AI-powered cyberattack.

Alicia Shapiro

Agentic Workflows

Google's New Gemini Agent Could Coordinate Business Work—Even With Claude

Oct 12, 2026

•

22 min read

Google's New Gemini Agent Could Coordinate Business Work—Even With Claude

Google isn't just building another AI coworker. It wants Gemini to become the system that organizes how businesses work—even when a competitor's AI is doing the job.

Alicia Shapiro

AI Governance

Meta’s Muse Can Look Like Human Traffic. Can Websites Trust AI Agents?

Oct 7, 2026

•

19 min read

Meta’s Muse Can Look Like Human Traffic. Can Websites Trust AI Agents?

An AI agent can shop, log into accounts and even complete security checks for a real customer—but what happens when the website has no idea it’s dealing with AI?

Alicia Shapiro

AI Safety and Security

Tavus Griffin: The Human-Like AI That 48% of People Thought Was Real

Oct 5, 2026

•

27 min read

Tavus Griffin: The Human-Like AI That 48% of People Thought Was Real

For most of the internet’s history, seeing a face react to you in real time was pretty good evidence that another person was there. That assumption may be about to change.

Alicia Shapiro

AI Safety and Security

Can AI Agents Be Contained? NVIDIA Moves AI Agent Security Outside the Model

Oct 2, 2026

•

16 min read

Can AI Agents Be Contained? NVIDIA Moves AI Agent Security Outside the Model

AI agents have always needed limits. The harder problem is making those limits hold when an increasingly capable agent can search for another way to complete its task.

Alicia Shapiro

AI Governance

OpenAI DevDay 2026: AI Agents Become Persistent Workers — What Should They Be Allowed to Do?

Sep 30, 2026

•

20 min read

OpenAI DevDay 2026: AI Agents Become Persistent Workers — What Should They Be Allowed to Do?

OpenAI’s newest agents are being built to keep working after you walk away — which makes the next question less about what AI can do and more about what it should be allowed to do.

Alicia Shapiro

AI Safety and Security

OpenAI Discloses 6 AI Misalignment Incidents: Can AI Safely Monitor AI?

Sep 18, 2026

•

30 min read

OpenAI Discloses 6 AI Misalignment Incidents: Can AI Safely Monitor AI?

Humans can't watch every move an autonomous AI agent makes. So researchers are turning to another AI to do it.

Alicia Shapiro

AI Agents

ChatGPT Work Website Sign-In Now Lets Tasks Run After You Leave

Aug 28, 2026

•

10 min read

ChatGPT Work Website Sign-In Now Lets Tasks Run After You Leave

What can ChatGPT Work finish after you sign in and step away—and how much account access does its cloud browser retain?

Alicia Shapiro

New AI Products and Features

Meta Launches Muse Code for Long AI Coding but Keeps Its Model Closed

Aug 7, 2026

•

16 min read

Meta Launches Muse Code for Long AI Coding but Keeps Its Model Closed

Would your engineering team accept a closed AI coding system to cut costs, or does long-term control matter more?

Alicia Shapiro

AI Voice

OpenAI Presence Gives Enterprises Control Over Voice and Chat Agents

Jul 24, 2026

•

10 min read

OpenAI Presence Gives Enterprises Control Over Voice and Chat Agents

What happens when a voice or chat agent’s correct answer or approved action becomes outdated after launch?

Alicia Shapiro

Agentic Workflows

Perplexity Study Shows AI Agents Expanding Knowledge Work

Jun 19, 2026

•

9 min read

Perplexity Study Shows AI Agents Expanding Knowledge Work

When agents can carry more of the workflow, the harder question is deciding what work is worth handing over.

Alicia Shapiro

Enterprise AI

JPMorgan Plans Long-Running AI Agents for Real Enterprise Workflows

Jun 17, 2026

•

7 min read

JPMorgan Plans Long-Running AI Agents for Real Enterprise Workflows

JPMorgan is preparing AI agents that can work for hours, raising a key enterprise question about when companies can trust agents with real workflows.

Alicia Shapiro

AI Governance

Microsoft Tests AI Evaluators for Trusted AI Agent Governance

Jun 12, 2026

•

12 min read

Microsoft Tests AI Evaluators for Trusted AI Agent Governance

If AI agents are being graded by AI, businesses need to know who checks the grader.

Alicia Shapiro

New AI Products and Features

Anthropic Launches Claude Opus 4.8 for More Reliable AI Agents

May 29, 2026

•

10 min read

Anthropic Launches Claude Opus 4.8 for More Reliable AI Agents

AI agents are getting more capable, but Claude Opus 4.8 puts the harder question in front of businesses: can they be trusted to work longer without making quiet mistakes?

Alicia Shapiro

Agentic Workflows

Claude Managed Agents Add Dreaming to Improve AI Agent Reliability

May 7, 2026

•

8 min read

Claude Managed Agents Add Dreaming to Improve AI Agent Reliability

AI agents are moving into longer, more complex work, and that raises a harder question: can teams trust how they remember, review and divide the work?

Alicia Shapiro

AI Safety and Security

World ID Expands Human Verification for AI Agents and Apps

Apr 24, 2026

•

13 min read

World ID Expands Human Verification for AI Agents and Apps

As AI agents begin acting on behalf of people, World ID is trying to answer a harder trust question: who is the real human behind the action?

Alicia Shapiro

AI Safety and Security

OpenAI Uses GPT-5.4 to Monitor AI Agents, Revealing Misalignment Risks

Apr 2, 2026

•

19 min read

OpenAI Uses GPT-5.4 to Monitor AI Agents, Revealing Misalignment Risks

OpenAI is using its most advanced AI to monitor its own coding agents in production, revealing what AI misalignment actually looks like in real-world systems.

Alicia Shapiro

AI Research

Microsoft Adds Multi-Model AI to Copilot Researcher to Improve Trust in Enterprise Research

Mar 31, 2026

•

16 min read

Microsoft Adds Multi-Model AI to Copilot Researcher to Improve Trust in Enterprise Research

Microsoft just gave its AI research agent the ability to question its own answers — using multiple frontier models to generate, review, and compare results before delivering them.

Alicia Shapiro

Future of Work

monday.com Launches Agentalent.ai, Letting Enterprises “Hire” AI Agents as Workforce Labor

Mar 27, 2026

•

11 min read

monday.com Launches Agentalent.ai, Letting Enterprises “Hire” AI Agents as Workforce Labor

What if hiring your next team member meant selecting an AI agent instead of a person?

Alicia Shapiro

Reasoning Models

ARC-AGI-3 Benchmark Shows Top AI Models Fail at General Reasoning, Scoring Below 1%

Mar 26, 2026

•

15 min read

ARC-AGI-3 Benchmark Shows Top AI Models Fail at General Reasoning, Scoring Below 1%

A new benchmark is forcing a hard question: are today’s most advanced AI models actually capable of reasoning—or just very good at pattern recognition?

Alicia Shapiro

AI Agents

Mar 22, 2026

•

17 min read

NVIDIA Launches Agent Toolkit for Enterprise AI Agents at Scale

AI is moving beyond generating answers—and starting to execute real work across enterprise systems.

Alicia Shapiro

AI Regulations

Meta Updates Moltbook Terms to Make Users Liable for AI Agent Actions

Mar 19, 2026

•

10 min read

Meta Updates Moltbook Terms to Make Users Liable for AI Agent Actions

Meta just made something very clear: if your AI agent acts, you’re the one accountable.

Alicia Shapiro

Future of Work

Anthropic Dispatch Lets Users Run AI Agents on Their Computers Remotely—Extending AI Beyond the Desktop

Mar 18, 2026

•

11 min read

Anthropic Dispatch Lets Users Run AI Agents on Their Computers Remotely—Extending AI Beyond the Desktop

What if your AI could keep working on your computer—even after you walk away?

Alicia Shapiro

Future of Work

Meta Brings AI Agents to Local Computers with Manus for Real-World Automation

Mar 17, 2026

•

14 min read

Meta Brings AI Agents to Local Computers with Manus for Real-World Automation

You can now tell an AI agent what to do from your phone—and it will complete the task on your computer at home.

Alicia Shapiro

Future of Work

Perplexity Launches “Perplexity Computer,” a Multi-Model AI System That Automates Complex Workflows

Mar 11, 2026

•

11 min read

Perplexity Launches “Perplexity Computer,” a Multi-Model AI System That Automates Complex Workflows

What if AI didn’t just answer questions—but ran entire workflows across multiple AI models and tools on your behalf?

Alicia Shapiro

AI Agents

Anthropic Warns Self-Replicating AI Could Spread to Your Smart Fridge

Google's New Gemini Agent Could Coordinate Business Work—Even With Claude

Meta’s Muse Can Look Like Human Traffic. Can Websites Trust AI Agents?

Tavus Griffin: The Human-Like AI That 48% of People Thought Was Real

Can AI Agents Be Contained? NVIDIA Moves AI Agent Security Outside the Model

OpenAI DevDay 2026: AI Agents Become Persistent Workers — What Should They Be Allowed to Do?

OpenAI Discloses 6 AI Misalignment Incidents: Can AI Safely Monitor AI?

ChatGPT Work Website Sign-In Now Lets Tasks Run After You Leave

Meta Launches Muse Code for Long AI Coding but Keeps Its Model Closed

OpenAI Presence Gives Enterprises Control Over Voice and Chat Agents

Perplexity Study Shows AI Agents Expanding Knowledge Work

JPMorgan Plans Long-Running AI Agents for Real Enterprise Workflows

Microsoft Tests AI Evaluators for Trusted AI Agent Governance

Anthropic Launches Claude Opus 4.8 for More Reliable AI Agents

Claude Managed Agents Add Dreaming to Improve AI Agent Reliability

World ID Expands Human Verification for AI Agents and Apps

OpenAI Uses GPT-5.4 to Monitor AI Agents, Revealing Misalignment Risks

Microsoft Adds Multi-Model AI to Copilot Researcher to Improve Trust in Enterprise Research

monday.com Launches Agentalent.ai, Letting Enterprises “Hire” AI Agents as Workforce Labor

ARC-AGI-3 Benchmark Shows Top AI Models Fail at General Reasoning, Scoring Below 1%

NVIDIA Launches Agent Toolkit for Enterprise AI Agents at Scale

Meta Updates Moltbook Terms to Make Users Liable for AI Agent Actions

Anthropic Dispatch Lets Users Run AI Agents on Their Computers Remotely—Extending AI Beyond the Desktop

Meta Brings AI Agents to Local Computers with Manus for Real-World Automation

Perplexity Launches “Perplexity Computer,” a Multi-Model AI System That Automates Complex Workflows

AiNews.com