17 April 2023

Good morning, and happy Monday!

Kicking off this week with a blog post by the UK Information Commissioner which I missed a couple of weeks ago, setting out 8 questions that Generative AI developers and users need to ask. It’s not great, to be honest, but might be a handy resource.

There are a couple of really interesting military stories, including one on how soldiers can now steer robot quadrupeds with brain signals, using a small sensor tucked behind their ears. (mad!) There’s also a piece on AI enabled military simulations for training purposes, which is interesting I think in terms of AI’s potential to assist in areas like the proportionality (collateral damage) assessment. There is also a piece on Ukraine’s use of autonomous naval drones, and an acknowledgement that AI is used for target recognition, although it is not clear from the piece whether this refers to visual identification, or target generation.

There are, as ever, a few pieces on Generative AI. Looking at Italy’s decision to ban ChatGPT, the use of generative AI to debug code on the fly (something I would have loved back when I was a programmer), and loads of chat on education. There is one story that could be a really interesting future line of research, looking at our unwillingness to engage with, or take instructions from, machines.

Today’s song is London.Paris.Berlin since I’ve been travelling a bit, and its lovely.

Chatham House, Recalibrating Assumptions on AI

Information Commissioner’s Office, Generative AI: eight questions that developers and users need to ask

Defense One, Soldiers Can Now Steer Robot Dogs with Brain Signals

Defense One, How One Simulation Maker is Adding AI, Drone Tactics

WIRED, Ukraine’s Quest for Homegrown AI Drones to Take on Russia

The Register, China reportedly let AI control a satellite, which then observed rivals India and Japan

Algorithm Watch, The algorithm that blew up Italy’s school system

Machine Intelligence Research Institute, Pausing AI Developments Isn’t Enough. We Need to Shut it All Down.

CNBC, Italy became the first Western country to ban ChatGPT. Here’s what other countries are doing

AI News, Italy will lift ChatGPT ban if OpenAI fixes privacy issues

ArsTechnica, Developer creates “regenerative” AI program that fixes bugs on the fly

ScienceNews, How ChatGPT and similar AI will disrupt education

The Register, Turns out people don’t like it when they suspect a machine’s talking to them

The Washington Post, We’re Not Ready to Be Diagnosed by ChatGPT

The Washington Post, Another Israeli spy app has been sold to 10 countries, researchers say

The New York Times, The Famous Black Hole Just Got Even Darker

The Washington Post, The AI bot has picked an answer for you. Hears how often its bad.

The New York Times, Security Robots. DigiDog. GPS Launchers. Welcome to New York

Subscribe to our weekly newsletter