Want to de-escalate protests on campus? Easy. Just send in a pack of military grade, drug-sniffing robot dogs that can be remotely controlled from anywhere in the country.

h/t Allegra Harpootlian

edtechmagazine.com/higher/article…

We are pleased to invite you to attend our online co-hosted event with the International Committee of the Red Cross on Artificial Intelligence in Military Decision Making. The event will launch two significant reports on the subject: geneva-academy.ch/event/all-even…

As tech companies get more desperate to generate revenue with AI, they’re going to make it harder to opt-out from using it. Anytime you want to search the web, write an email, or drop a bomb you’ll have to decline an evermore forceful “offer” to let AI do it for you.

When the War Machine Decides: Algorithms, Secrets, and Accountability in Modern Conflict

Deep thanks to Arthur Holland Michel and Carnegie Council
for inviting me to speak about the future of AI in war.

Listen: carnegiecouncil.org/media/podcast/…

This was so much fun! A recording will be up soon 👀

Tomorrow, virtually at 6pm ET! Our very own Research Director Eleni Manis will be joining Carnegie Council's Arthur Holland Michel and Harvard Kennedy School's Chris Gilliard. Together, they will discuss the ethics of data fusion. Register using this link: info.carnegiecouncil.org/impact-of-data…

On Feb 13 I’ll be talking to two of the brightest stars tracking surveillance tech about an issue that you need to start caring about asap. Tune in!

i am gutted by how many talented people in journalism have lost their jobs today, yesterday, last week, etc. the system is so broken and not every journalist will be able to make a living off of newsletters. there needs to be a larger solution

Jeff Bercovici Lindsay Blakely Carly Olson Not sure what's next for me—if you've got an idea, my inbox is always open. Opening for a tech columnist? Writer? Editor? Something in academia, advocacy or teaching? A big ticket offer from a giant corporation that makes me question all my values? HMU

briancmerchant at gmail

Here we go...

'OpenAI changes policy to allow military applications'

techcrunch.com/2024/01/12/ope…

OpenAI: 'it would be impossible to train today’s leading AI models without using copyrighted materials.”

Is this a real legal argument that has ever actually worked?

theguardian.com/technology/202…

In light of the attempted drone boat attack in the Red Sea, re-upping this story I did for The Economist on the threat that the technology poses—not just to navies but to everyone who does business on the sea.

economist.com/science-and-te…

And there it is folks. Some of your favorite AI was trained on a dataset that contains thousands of instances of child sexual abuse material. Confirms what many of us long suspected.

What we don't know is:

❌ How many of the more than 22,000 targets in this conflict were generated by the Gospel.

❌ How many of those targets were rejected by human analysts.

❌ The BDAs of the sites the system did elect to strike (i.e. is it effective?)

Re-upping this report I wrote with ARES for the ICRC on explosive weapons - how they work, how targeting is done, what effects do weapons have, fragment numbers, lethality, etc - as we are seeing bombs falling in #Gaza and rockets in #Israel : icrc.org/en/document/ex…

Friendly reminder that any civilian-saving benefits a smart AI targeting tool *might* proffer are likely eviscerated if you use big dumb bombs in the corresponding strikes.

New: Tesla is recalling nearly all 2 million of its cars on US roads to limit use of its Autopilot feature after a review of nearly 1,000 crashes in which it was engaged. The recall was disclosed in a letter to Tesla posted by the National Highway Traffic Safety Administration.

Utterly bonkers finding.