April 19, 2023

  • Deepfakes

    Schumacher family planning legal action over AI ‘interview’

    Source, 04/19/23. Michael Schumacher’s family plans legal action against a German magazine for publishing an AI-generated “interview” with the seven-time Formula One champion. Schumacher has not been seen in public since his 2013 skiing accident, and the family has guarded his privacy, limiting access to information about his condition. Schumacher’s son, Mick, is currently the reserve driver for Mercedes in Formula One. READ MORE

  • Stability AI

    Stability AI releases ChatGPT-like language models

    TechCrunch, 04/19/23. Stability AI has open-sourced StableLM, a suite of text-generating AI models available on GitHub and Hugging Face. The models can generate both code and text and were trained on a custom dataset. While the release of open source models has been criticized, Stability AI argues that open-sourcing promotes transparency and fosters trust. However, the company is under legal pressure and has been criticized for its tools’ use in generating deepfakes and graphic depictions of violence. Stability AI is under pressure to monetize its efforts and has hinted at plans to IPO. READ MORE

  • Jail Break

    People Are Using A ‘Grandma Exploit’ To Break AI

    Kotaku, 04/19/23. Artificially intelligent text generation tools like ChatGPT and Bing chatbot are causing concern, but users are finding new and amusing ways to exploit them. For example, they are asking AI to pretend to be deceased grandparents who knew how to make incendiary weapons. This highlights both the potential for negative implications and the creative potential of the technology. Developers need to mitigate potential risks while embracing the opportunities presented by these emerging tools. READ MORE

  • Google

    Researchers Accused Google And “60 Minutes” Of Spreading AI “Disinformation”

    BuzzFeedNews, 04/19/23. AI researchers criticized CBS and Google for overhyping AI in a 60 Minutes interview with Google CEO Sundar Pichai. The segment featured a program claimed to have learned a language it had never seen before all by itself, but researchers pointed out it was trained on Bengali. They called for greater accountability from tech companies in the development and deployment of AI technology. READ MORE

  • The Pause

    Military Tech Execs Tell Congress an AI Pause Is ‘Close to Impossible’

    GIZMODO, 04/19/23. Experts warned against pausing AI development due to concerns that China may capitalize on the lull to surpass the US. Instead, they recommended smart regulations and increased spending on capabilities to “terrify our adversaries.” They also suggested offensive actions, like “data poisoning,” to limit other countries’ ability to develop superior AI systems. READ MORE