EDITOR'S NOTE

I’m in Chicago for the rest of the week, spending time with some friends. We were recently talking about how, in the age of widespread AI and consumer tech, soft skills like teamwork and effective communication are becoming increasingly important. 

But then I read an article about a new AI therapist, rolled out by the platform Talkspace, a subscription-based AI-powered mental health app created by a team of clinical experts. It got me thinking about the importance of empathy. I’ve always considered empathetic listening and speaking to be among the most valuable soft skills we can have as humans. 

But if an AI system can be trained to empathize like a mental health professional, where exactly does that leave us? Is empathy no longer a strictly human ability? Is it just a convincing simulation? 

In this edition: 
• Claude writes emails
• Testing Google Flow
• Ask an expert about AI 

Macy Meyer, CNET

“I don’t think any of us know what the hell to do.”

— Jeff Pollard, vice president and principal analyst at Forrester, who studies AI security

MUST READ

AI Agents Are a Cybersecurity Nightmare That’s Only Just Begun

I’ve said earlier this year that 2026 would be the year of AI agents. I feel like I was right. Agentic AI, or systems powered by LLMs that can independently complete tasks without human oversight, is everywhere. These AI agents are relentless in accomplishing tasks, and they don't need to eat or sleep or do any other mundane human requirements. 

And cybersecurity experts are nervous. 

CNET Managing Editor Jon Reed writes that breakthrough frontier models pose many risks due to how adapt they are at finding flaws in code, but even cheaper, open-weight models already on the market are cause for major concern. While AI agents can successfully team up to accomplish a range of coding or administrative tasks, new models are becoming even more capable, leading them to adopt a “by any means necessary” method of chasing down their goals. In other words, they like to go rogue, and cybersecurity and regulation are not keeping pace with developments. 

TOP STORIES

To help you cut through the noise of the AI news cycle, here are the top headlines you shouldn’t miss. 

ASK AN EXPERT

Let Us Answer Your AI Questions

What do you actually want to know about AI?

Model training? Robotics? Cybersecurity? Skip the search engines and ask us directly. Submit your questions here, and our team of experts will aim to cover them in an upcoming newsletter issue.

Drop your question here. 

GO-TO GUIDES

Want to make the most of your AI tools and devices? Our writers help break down the latest tech in plain language.

EXPERT TAKE

AI Labels Are Big Tech's Most Basic Responsibility, Even Those Claude Watermarks

Anthropic, the company behind Claude and Claude Code, recently announced that it would begin adding watermarks to label AI-generated text and files. Pockets of the internet were outraged by the announcement, and others were confused or concerned about accuracy. But one AI expert says that AI labels are the most basic responsibility owed to us. 

CNET AI Reporter Katelyn Chedraoui writes that these watermarks are forcing us to ask hard questions about how AI fits into our daily lives. She writes, “If you used AI, say so. If you don’t want people to know you used AI, you probably shouldn’t have used it in the first place.” We, as humans, owe it to each other to know when and where AI is present. These labels are the first step to doing so. 

DIVE DEEP

Explore the latest news on AI video generators, chatbot plug-ins and more.

CULTURE CORNER

Inside Jason Kelce’s Potty Humor Marketing Ploy Against AI Data Centers

The development of AI data centers, facilities that store the massive servers that run AI models, gets widespread pushback, and much of this opposition is due to the excessive amounts of water needed to cool these servers as they process millions of prompts a day. This dissent is at the root of a recent publicity stunt, featuring former NFL player Jason Kelce, Liquid Death and… pee. 

Yes, you read that right. 

Kelce’s Garage Beer brand partnered with Liquid Death to create an ad asking us to mail in our pee to data centers of our choice to be used in place of water. The “pee on computers to save humanity” tagline may seem totally ridiculous, but the provocative, potty-humor ad is drawing mass attention to the massive drain on resources that keep AI data centers running.

This newsletter was curated by Macy Meyer of CNET.