– AI offended an actress and lost its voice
– Sony doesn’t want AI to listen to its music
– Researchers explore the “black box” to decode the AI mind
This week in AI news, Scarlett Johansson was offended by the voice of GPT-4o, which sounded eerily like hers. Sony Music Group warned companies that its music is off-limits for AI training, raising questions about the use of copyrighted data. Researchers at the University of Texas found a way to train models without using copyrighted work, while Bollywood sees a growing trend of political AI deep fakes.
Microsoft unveiled more AI-powered work automation tools, while leading AI companies agreed to voluntary safety commitments ahead of the AI summit in Seoul. Google published its Frontier Safety Framework to mitigate AI risks, acknowledging potential dangers.
Yann LeCun disagrees with the idea of an extinction threat from AI, stating that current language models are no more dangerous than having access to a library. In the gaming world, FRVR uses generative AI to create games from natural language, allowing users to easily make their own games on their platform.
AI events like the City Week conference in London and the Enterprise Generative AI Summit in Silicon Valley discussed the impact of AI on industries. The COMEX Global Technology Show in the Middle East offers a glimpse into a future shaped by AI, VR, and blockchain. Other interesting AI stories from the week were highlighted as well.