Microsoft belongs to the court to abuse AI instruments

Microsoft is trying to show up to AI security change the claim To open the four developers last year, this went out of a girl in AI to create famous DEEPFakhes. The company filed a lawsuit in December and the court decision, which allows Microsoft to seize a website related to the operation, determined individuals.

Arian Yadaria and Iran Fiz; Alan is the kingdom of Krysicak drago; Hong Kong, Kong Kong, Vietnam’s Asakari book book.

Microsoft said there were others identified as in the scheme, but he does not want to call not to intervene in the ongoing investigation. According to Microsoft, according to Microsoft, it was able to “jailbreak” to create access to its generative AI instruments and create types of images they want. The group then included others who used to create a deep deposit nude of celebrities between other abuse.

After the lawsuit filed a lawsuit and seized the group’s website, Microsoft said the suspects went to panic. “The immediate response of actors to seize this website and not opening the legitimate documents in January, in some cases, the fingers of the group’s fingers lead to each other and indicate each other,” he said.

Celebrities, including taylor swiftIt is often often a pornography that takes a real person’s face and confidently superior to a naked body. In January 2024, Microsoft had to update Microsoft text-image models after visualizing the false images of fake footage visible on the Internet. Generative AI makes it incredibly easy to create images with small technical abilities – already caused the epidemic of secondary schools in the United States Experiencing deepfake scandals. The latest stories The victims of depth Creating images, because digitally occur, but to use the targets, how to use enough people to make it afraid and to do there, how to use real-world damage to recognize enough people to do.

The AI ​​community has an argument that is a real or more concerns of safety and concerns to help Openai influence large players and sell their products Do not prefer true power generative artificial intelligence. A camp claim that the EU models prevented the worst violations by restricting the capacity of users to turn off security controls; Open-source campers are possible to prepare free models to speed up and develop the sector and improve, without hindering the innovation. In both cases, all feel like a slight distraction from the nearest danger, which fills the AI ​​Internet with inaccurate information and sliding content.

Many fear related to AI feels in nature and hypothetical and are likely to be in a good place to abuse the EU’s deep, as a generative AI is in a good enough place to get their agency. This is a way that this abuse can be solved so far. Persons already have a detention decision in the United States Used AI to create a depth of minorsand and and There is no fake action The Congress was introduced last year, and it would be a crime to create pictures based on the likeness of someone. The United Kingdom already punishes the distribution of Deepfake porn and will soon be a crime produce. Australia recently committed his creativity and sharing until the consensual depth.