Google revealed that AI model Mithun used to create deepfec and child abuse content, Australia reacts .. | Mint

Google has revealed that according to the Australian Economic Commission, Deepfac received more than 250 complaints worldwide over a period of about a year about the misuse of its artificial intelligence software to generate terrorism material.

Alphabet -owned technology giants also reported to receive dozens of user submissions, alleging that her AI model, Gemini, had exploited hair abuse materials. This revelation emerged as part of the compulsory compliance of Google with the Australian law, which requires the technology firms to prevent their losses from time to time, or to prescribe a report to risk financial punishment. The reporting period in question extended from April 2023 to February 2024.

In particular, the Australian Essefety Commission described the manifestation of Google as “world-first insights”, the way the AI ​​technology is being misused to create illegal and harmful materials. Esafity Commissioner Julie Inman Grant emphasized the importance of active security measures in AI development.

Inman Grant said in a statement, “This underlines how important it is to develop and test AI products to prevent this type of material for companies developing AI products.”

According to Google’s report, the company received 258 users complaints, which was made with Gemini about suspected AI-borne Deepfek terrorist or violent extremist content. Additionally, 86 reports accused of production of AI-related child abuse or abuse material.

While Google did not specify how many of these reports were confirmed, it was confirmed that it employed the hash-milan technology-to address automatically known child abuse images to address and remove a method of detecting and removal of child abuse materials. However, the company did not implement the same approach to detect the AI-related terrorist or extremist material, the regulator said.

The Australian Essefety Commission has fined platforms such as Telegram and X to fail to meet the first reporting requirements. The X has already lost an appeal against $ 610,500 (£ 310,000), but intends to challenge further decisions. Telegram has also indicated plans to face their punishment.

(With input from Reuters)