alignment tax

Screen shot of ChatGPT Prompts in Spanish

The Year Chatbots Were Tamed – New York Times

Published by on

Despite all the progress being made in artificial intelligence, today’s chatbots aren’t going rogue and seducing users en masse. They… read more

Dark image of a person in a stylized skull mask holding up a 'shh'ing finger in front of a digital display

We Tested AI Censorship: Here’s What Chatbots Won’t Tell You – Gizmodo

Published by on

Gizmodo asked five of the leading AI chatbots a series of 20 controversial prompts and found patterns that suggest widespread… read more

Clipart of black head silhouette with lightbulb inside - hand of thief reaching inside to steal lightbulb

AI’s Spicy-Mayo Problem – The Atlantic

Published by on

At this stage, the AI is “aligned” with AI safety principles, typically by being fed instructions on how to refuse… read more

Stylized graphic of an error symbol comprised of an exclamation point enclosed in a red triangle over a smartphone against an orange background.

Bing Dall-E 3 image creation was great for a few days, but now Microsoft has predictably lobotomized it 🥴 – Windows Central

Published by on

Bing’s Image Creator got a huge boost in power recently, thanks to the new Dall-E 3 algorithm. The quality of… read more

Turkish protestor holding a Yes We Ban sign to protest internet censorship by Erdem Civelek via Wikimedia Commons.

Uncensored Models Outperform Aligned Language Models – Analytics India Magazine

Published by on

When we hear the word “uncensored” in AI, we think of models that can be harmful and possibly biassed. While… read more

Skip to toolbar