Skip to content
TheNews SyncTheNews Sync

Searches the headlines of the last three days, from every outlet we follow.

Menu

Reading

Text size
Saved on this device. Nothing about you is sent to us.

Account

Your name on comments, on any device
International
International
TechUpdated

Microsoft says AI agents are like insider threats as ‘frontier models can’t be trusted’

From TechCrunch ·
Headline from TechCrunch
1 outlet · 1 report · latest EN 1English 1 · all privateOwners: all private.
DiscussionWhatsApp

Microsoft wants enterprises to closely watch the agents employees use with systems not controlled by AI labs like OpenAI and Anthropic.

There’s no combined summary for this story yet — read the outlets’ reports below.

Your take

Be the first to react

Discussion

Readers' views on this story
Join the discussion

Story alerts

Get an email when an outlet reports something new on this story.

Discussion

Readers’ views on this story · published as written
No account needed ·
Community rules

Read the originalsRead the original reports 1

1 reportShowing 1 of 1

Keep reading

Up next · Weather & disasterSpain floods kill four as torrential rain inundates eastern regionsFrance 24
Politics · Economy & businessSri Lanka hit by islandwide power outage12 outletsCrime & police · PoliticsBalangoda Kassapa Thera hospitalised during satyagraha8 outletsForeign ties · Economy & businessChina Merchants logistics hub project off; SLPA agrees to $21.9m settlement deal7 outlets

Sign in

Optional. No password — we email you a code. Signed in, your alerts start at once and your comments carry your name on any device.