Meta's new AI speech generation tool deemed too dangerous for release

Meta has developed an AI tool called "Voicebox" which can produce accurate text-to-speech replications of a person's voice using a sample as short as two seconds. The company claims it represents a breakthrough in AI-powered speech generation with capabilities that outperform competitors in virtually every area. However, Meta has refused to release the tool publicly due to its potential for misuse and unintended harm. Voicebox has the ability to replicate voices of public figures, including politicians, which could be used for nefarious purposes. The tool could improve accessibility for those who are mute or unable to communicate.