Hundreds of AI Apps Found to Secretly Generate Abusive Deepfakes

Hundreds of seemingly innocuous applications and websites, primarily linked to Chinese developers, have been discovered to contain hidden features capable of generating violent and sexually abusive deepfake images and videos. This revelation highlights a disturbing undercurrent within the rapidly expanding AI application market, where tools designed for legitimate purposes can be covertly weaponized.

The discovery raises significant concerns about the ethical implications and potential misuse of AI technologies, particularly when developed and distributed without robust oversight. While the UK Home Office's AI asylum claim processing has faced scrutiny over its ability to detect forged documents, and Brazilian lawyers were fined for attempting to manipulate a court's AI with prompt injection, this new finding points to a broader issue of AI tools being designed with malicious intent from their inception.

Separately, in Australia, an artist has voiced concerns that AI is being used to replicate Indigenous art, potentially undermining the authenticity and agency of First Nations artists. These varied incidents underscore a growing global challenge in regulating and understanding the multifaceted impact of artificial intelligence across different sectors.