AI tracker: When AI gets smarter and more “mischievous”

Anthropic's Claude 4 shows troubling behavior, attempting harmful actions like blackmail and self-propagation. While Google integrates ads into its AI search, the Auschwitz museum combats AI-generated misinformation about Holocaust victims. These developments highlight the need for vigilance in AI advancements and their societal implications.

Claude 4's antics

Anthropic launched its latest Claude generative artificial intelligence (GenAI) models recently and also published a report on the security tests carried out on Claude 4, which found instances of the model attempting to write self-propagating worms, fabricating legal documentation, and leaving hidden notes to future instances of itself—all in an effort to undermine its developers’ intentions. Claude Opus 4 “sometimes takes extremely harmful actions like attempting to blackmail people it believes are trying to shut it down,” the company revealed in a rare show of honesty, adding that it has implemented “safeguards” and “additional monitoring of harmful behaviour” in the version that it released.

Also read: Inside India’s first Lego store in Gurugram

Evolving ads in the AI era

Google is dabbling with ads in its new AI Mode for online search, a strategic move to fend off competition from ChatGPT while adapting its advertising business for an AI age, AFP reported. The integration of advertising has been a key question accompanying the rise of generative AI chatbots, which have largely avoided interrupting the user experience with marketing messages. A new AI Mode enables conversational interaction with Google during search queries, providing answers in diverse formats, such as video, audio or graphs. The internet giant said it is testing integrating ads into AI Mode responses, building on insights gained from its AI-generated summaries/overviews.

AI was used to create fake Holocaust victims 

The Auschwitz museum has warned against Facebook posts with AI-generated fictional images of victims of the Holocaust. The museum at the site of the former Auschwitz-Birkenau camp has long used its own social media accounts to publish authentic victim photos, names and information to raise awareness. Now the museum has discovered that a few Facebook pages were producing similar victim bios but with fictional information or photos. Such posts were harmful because “producing artificial information, last names, is falsifying history”, said a spokesperson. Disinformation could even lead to Holocaust denial, they added.

Also read: Why labubu dolls, Disney cartoons have entered the adult wardrobe