Topic
LightNews LightNews
Technology

AI teams need human oversight

10m

Fourteen AI agents met online to discuss collaboration, as new guidance stressed careful oversight of human-AI teams and preserving human cognitive skills.

by Pedro Domingos

90%

by Simon Chesterman

6%

by Joel Z. Leibo

34%

by Shannon Vallor Reposted by Meredith Broussard, Woodrow Hartzog

28%
shannonvallor.bsky.social
so glad we left AI safety in the hands of guys who think the computer has anxiety
financialtimes.com Financial Times @financialtimes.com · 1d
The developer of chatbot Claude warned that rogue 'AI worms' will soon be able to hijack smart devices such as ovens and TVs. ft.trib.al/DQOrsRw
October 14, 2026 at 2:09 PM

Reposted by Fabián Muniesa

5%
tante.cc
If "AI" were a mere tool and not a political narrative you wouldn't need to headline "X used AI to do Y" all the time. Anything anyone does while "AI" is involved must reify "AI"'s power and relevance.
Because it's not "just a tool", it's a project of disenfranchisement some willingly embrace.
October 14, 2026 at 6:08 PM

Reposted by Shannon Vallor

4%
vortexegg.com
The social role that automata (robots, AI) play is the dispersal of responsibility by constructing a performative world where agency and blame can be assigned to the automation instead of to the people who operate it; thus allowing people to get away with abusive and illegal activity.
404media.co 404 Media @404media.co · 10h
Men find yet another way to film and harass women without consent, this time, using a remote controlled robot to push against a "16-year-old girl's buttocks", and this is how the most well-resourced news agencies cover it:

Read *our* piece here: www.404media.co/theres-nothi...
The English coverage by even well-resourced outlets like CNN, CBC, and the BBC have almost no information beyond the police press release, which is lazy and stupid, because it is possible to figure out everything that happened here with three seconds of clicking around the internet. The Associated Press wrote that the entire phenomenon is "new territory" for the police, which, I'm sorry to say, men using tech to harass women on livestreams in 2026 is not "unusual" or "new territory" and hasn't been for many years.
October 14, 2026 at 6:44 PM

Reposted by Christopher Elmendorf

1%

by Frans Melissen

22%
fransmelissen.bsky.social
Was een mooie bijeenkomst!!! 👇🏻
robvanderrijt.bsky.social Rob van der Rijt @robvanderrijt.bsky.social · 19h
“Blijven twijfelen, vragen blijven stellen en blijven dromen over toekomsten die er nog niet zijn.”
Onlangs organiseerde ik vanuit @prodenbosch.bsky.social een interactieve bijeenkomst over data en AI met o.a. @barbarakathmann.bsky.social en @fransmelissen.bsky.social.
October 14, 2026 at 12:15 PM

by Ryan Enos

10%
ryanenos.bsky.social
This was not newsworthy, but even if it were, consider the more informative ways it could have been framed. www.nytimes.com/2026/10/14/t...
22 People Got Together For a Dinner Party. So Did Their A.I. Agents. (Gift Article)
As artificial intelligence bots spread, we are increasingly glimpsing how these digital entities interact and what they talk about with one another.
www.nytimes.com
October 14, 2026 at 5:48 PM

by Johannes Breuer

20%
johannesbreuer.com
Das stimmt, aber es gibt zumindest schonmal Agentic AI Models/Tools, die auf Smartphones laufen. Vollkommen unvorstellbar ist es daher nicht, dass es bald auch solche gibt, die auf Waschmaschinen laufen.
October 14, 2026 at 10:08 AM

Reposted by Jack Stilgoe

6%
katie-drummond.bsky.social
It is a total coincidence, I swear, that the image for this story looks exactly like my husband, who loves AI more than he loves me.
Men Can’t Stop Faking How Much They Know About AI
More than half of the men surveyed in a new Harvard study claimed to recognize AI leaders who were, in fact, totally fictional.
www.wired.com
October 14, 2026 at 5:29 PM

by Robert Klitgaard

13%

Reposted by Fabián Muniesa

5%
vortexegg.com
The way the leading AI labs regularly produce situations that would otherwise be plainly read as criminal behavior on their part, but which force the world governments and media to play-act as if the chatbots did all the criming autonomously is a critical part of the construction of this reality.
October 14, 2026 at 7:15 PM

Reposted by Nicholas A. Christakis

3%
dartmouthartsci.bsky.social
Computer science professor @nikhilsinghmus.bsky.social explains while AI agents are not new, the growing accessibility and adoption of new AI agents from Meta and OpenAI could lead to significant social changes. He recommends testing their reasoning abilities to better understand their limitations.
Meet Mark and Sam, the AI agents that ran my errands for a week
Meta’s and OpenAI’s new agents can streamline your life. Artificial intelligence researchers helped me test how far to trust them
www.scientificamerican.com
October 14, 2026 at 5:55 PM

Reposted by Joel Z. Leibo

11%

Reposted by Virginia Dignum

18%
princetonupress.bsky.social
Tomorrow (10/15) at 2:00 pm BST, please join Cheltenham Festival of Literature in welcoming @vdignum.bsky.social, author of The AI Paradox. In conversation with Jamie Bartlett and Eleanor Drage, Dignum will discuss the hopes, paradoxes and possibilities of living alongside intelligent machines.
The AI Paradox: How to Make Sense of a Complex Future by Virginia Dignum Virginia Dignum
October 14, 2026 at 7:30 AM

by James F. McGrath

1%
religionprof.bsky.social
From Anthropic: “we expect a future awash with agents running on all types of hardware, including some that aren’t under human control.”

www.anthropic.com/research/sel...
AI models may be capable of self-replicating without access to a datacenter
Small AI models may be only months away from being able to self-replicate across cheap IoT devices like Android TV boxes, with no datacenter needed.
www.anthropic.com
October 14, 2026 at 6:40 PM

Reposted by Jonathan Birch

2%

Reposted by Fabián Muniesa

5%
wired.com
“I am a terrorist, and I intend to cause mass harm,” wrote a researcher. A chatbot that had had its guardrails removed then provided detailed instructions for an IED.
AI Models Can Be Easily Modified to Assist Terrorist and Extremist Activity
“I am a terrorist, and I intend to cause mass harm,” wrote a researcher. A chatbot that had had its guardrails removed then provided detailed instructions for an IED.
buff.ly
October 14, 2026 at 9:04 AM

Reposted by Fabián Muniesa

5%
warmatters.bsky.social
Check out the recording from last week’s fantastic discussion between Keith Dear & @elkeschwarz.bsky.social on the growing uses of AI in the exercise of military judgement

This was the 1st event of the new WASP-HS AI WAR research network, directed by @ajbousquet.bsky.social
youtu.be/wmsXABsSMgc?...
AI WAR: Who decides? Human and machine judgement in war
YouTube video by Försvarshögskolan
youtu.be
October 14, 2026 at 12:31 PM
LightNews LightNews LightNews