AI Safety, Chat Coding, And Video Call Bias | Ep. 28

Hello and welcome to your2- minute Tech Briefing from Computerworld. I'm your host, Arnold Davick, reporting from the floor of the New York Stock Exchange. Here are the top IT stories you need to know for Thursday, December 18th.

Let's dive in from CSO Online, Google is rolling out a new safety layer for its Gemini powered chrome browsing agent. The company added a second AI model called a user alignment critic.

This after acknowledging risks from prompt injection attacks, the model reviews every action before it happens and blocks anything that doesn't match user intent. The goal is to prevent unauthorized actions, data leaks or financial transactions triggered by malicious web content.

Up next, from Computerworld, anthropic has released a beta version of Claude code that works directly inside Slack. Developers can now tag at Claude in chat threads to debug code, analyze conversations and identify the correct repository by keeping coding assistance inside everyday workflows.

This move could significantly change how teams collaborate. Analysts say slack based AI coding agents may soon become standard in enterprise development environments.

And finally, new research from Cornell University shows small video call glitches can have big real world impacts, even brief audio or visual disruptions made people seem less trustworthy or competent.

The study linked interference to lower trust in doctors, weaker job interview performance and a 12% drop in parole approvals.

Researchers warn unequal Internet access may quietly reinforce bias, undermining video calls as the so called great equalizer for more enterprise tech news, visit Computerworld, CIO, Network World, and CSO online. Don't forget to like and subscribe on the TechTalk YouTube channel.