Google DeepMind has unveiled Genie 4, an updated world model that generates photorealistic video alongside matching human voices and realistic motion. The model supports real-time steering through natural language and includes Gemini-driven reasoning for complex events. Safety guardrails and Google's SynthID watermark are built in, and early testers include video game developers prototyping with natural language prompts.
Google plans to roll out Genie 4 first to trusted testers, then to Google AI Ultra subscribers and partners. Researchers are already exploring uses in animal behavior studies, disease spread modeling, filmmaking pre-visualization, and robot training scenarios. The model improves on Genie 3 with better physics, efficiency, and scalability.
In a separate development, Anthropic's Claude Duet topped a comparison test of six major AI agents, including Meta's Muse, Google's Bard, and Microsoft's Turing-NLG. Claude Duet stood out because it can learn and improve over time rather than delivering fixed answers, performing best across tasks like writing, translation, and question answering.
Dell Technologies is connecting supply chain security to the growing risks of agentic AI, noting that AI agents acting independently are harder to predict and control. Dell recommends a layered security approach covering hardware, runtime, data, and applications, and is working with Nvidia to address deployment challenges.
Meanwhile, several other developments are shaping the AI conversation. Trump announced a plan to ban the term "Artificial Intelligence" in official documents, preferring "Super Intelligence," with an executive order directing federal agencies to adopt the change. A student report found that many Russell Group universities are not adequately teaching AI skills, with access to tools often limited by paywalls. RTVE launched Telerín, a free AI tool for exploring Spanish Television history, on October 27 to mark TVE's 70th anniversary. Elon Musk said he has "cyber psychosis" amid growing discussion of AI burnout, while Apollo Global Management noted AI has delivered little financial return outside of tech. Texas voters are pushing back against AI data centre construction in their communities ahead of the 2026 midterm elections.
Key Takeaways
- Google DeepMind launched Genie 4, a world model generating photorealistic video, human voices, and realistic motion with Gemini-driven reasoning and SynthID watermarking.
- Genie 4 will roll out first to trusted testers, then to Google AI Ultra subscribers and partners, with applications in gaming, filmmaking, and research.
- Anthropic's Claude Duet outperformed Meta's Muse, Google's Bard, and Microsoft's Turing-NLG in an AI agent comparison test, thanks to its ability to learn over time.
- Dell Technologies is partnering with Nvidia to address agentic AI security risks, recommending layered protection across hardware, runtime, data, and applications.
- Trump plans to ban the term "Artificial Intelligence" in official documents, preferring "Super Intelligence," per an executive order dated September 29.
- A student report found many Russell Group universities lack adequate AI teaching, with some tools only available behind paywalls.
- RTVE launched Telerín, a free AI tool for exploring Spanish Television history, on October 27 marking TVE's 70th anniversary.
- Elon Musk cited "cyber psychosis" amid AI burnout discussions; Apollo Global Management reported limited financial returns from AI outside tech.
- Texas voters are opposing AI data centre construction in their communities ahead of the 2026 midterm elections.
- Computational neuroscientist David Sussillo noted AI may develop deep theories before neuroscience does, as the two fields once shared common ground but have diverged.
DeepMind Genie 4 adds audio and realistic people to simulations
Google DeepMind released Genie 4, an updated world model that adds audio and more realistic human characters. The model improves on Genie 3 with better physics, efficiency, and scalability. Researchers are already using it to study topics like animal behavior and disease spread. Google plans to make it available to researchers and developers in the coming months.
Google Genie 4 AI world model generates human voices and video
Google DeepMind launched Genie 4 as a limited research preview that generates photorealistic video and human voices. The model simulates realistic human motion, facial expressions, and interactive dialogue. It includes safety guardrails and will be tagged with Google's SynthID watermark. Early testers include video game developers building prototypes with natural language prompts.
Google launches Genie 4 realistic interactive world simulation model
Google DeepMind announced Genie 4, a multimodal world model that generates matching audio alongside visuals. It supports real-time steering through natural language, improved physics, human motion, and Gemini-driven reasoning for complex events. Access will start with trusted testers, then expand to Google AI Ultra subscribers and partners. Industries like filmmaking, gaming, and education are already experimenting with it.
Google's Genie 4 AI model creates more realistic experiences
Google announced Genie 4, a world model offering more realism and generated audio for interactive worlds. It is launching first as a research preview for trusted testers. Google says it could help with filmmaking pre-visualization, game prototyping, and robot training scenarios. The model will come to Google AI Ultra subscribers soon.
Trump says he will ban the term Artificial Intelligence
Donald Trump said he plans to ban the use of the term Artificial Intelligence in official documents. His preferred name is Super Intelligence, which he promoted during a White House meeting with tech executives. An executive order from September 29 directs federal agencies to use that term instead. The term AI has been in use for decades despite the proposed change.
Trump announces ban on term artificial intelligence
US President Trump stated that the country will ban the term artificial intelligence. He instead promoted the phrase super intelligence. The announcement was made in Washington DC on October 27, 2026.
Russell Group students report little AI teaching on courses
A student report found that many Russell Group universities are not adequately teaching students how to use AI. Training and access to AI tools vary widely between courses. Some AI features are only available behind paywalls. Experts stress the need for ethical and effective AI training in higher education.
RTVE launches free AI tool Telerín for TVE history
RTVE launched Telerín, a free AI tool that lets users explore early years of Spanish Television. It launched on October 27 to mark TVE's 70th anniversary. Users can search with natural language questions. The tool includes documents, photos, videos, and magazine archives from the 1950s and 1960s.
Elon Musk says he has cyber psychosis amid AI debate
Elon Musk said he has a little bit of cyber psychosis. The term has become popular in the tech world. The article discusses signs of AI burnout, including heavy chatbot use and new habits like smoking. Industry figures like Sam Altman and Marissa Mayer are also mentioned. Apollo Global Management noted AI has delivered little financial return outside of tech.
Neuroscience gave AI its start, what happens now
David Sussillo, a computational neuroscientist, discussed how neuroscience and AI are connected. AI originally grew from studies of the brain in the 1940s. The two fields once shared common ground but have since diverged. Sussillo worked at Google Brain, Meta Reality Labs, and Stanford University. He says AI may develop deep theories before neuroscience does.
Texas voters push back against AI data centre construction
Voters in the United States are united in opposing AI data centres in their communities. Jackson Proskow reported from Red Oak, Texas, about this backlash against big tech. This issue is gaining attention ahead of the 2026 midterm elections. Many people are concerned about the impact of these facilities on their local areas.
Philosopher questions if humans and AI truly understand each other
This article explores whether machines can truly understand the world. Philosopher Elizabeth Anscombe studied how language shapes reality. Joseph O. Chapa, who holds a Ph.D. in philosophy from the University of Oxford, writes about technology, ethics, and war. He is a research fellow at the University of Massachusetts-Boston's Applied Ethics Center. His upcoming book is called AI For The Rest of Us.
Anthropic's Claude Duet wins AI agent comparison test
A writer tested six major AI agents including Meta's Muse, Google's Bard, and Microsoft's Turing-NLG. The agents were challenged with tasks like answering questions, writing stories, and translating text. Anthropic's Claude Duet performed best overall across all tasks. It stood out because it can learn and improve over time rather than giving fixed answers.
Dell connects supply chain security to agentic AI risks
Dell Technologies says supply chain security is critical as companies bring AI into core business operations. AI agents can act independently, making their behavior harder to predict and control. Dell recommends a layered security approach covering hardware, runtime, data, and applications. The company also works with Nvidia to address challenges in agentic AI deployment.
Sources
- DeepMind's Genie 4 world model adds audio and more realistic people to simulations
- Google's New AI World Model Can Generate Human Voices, Not Just Video
- Google Launches Genie 4, An Even More Realistic World Simulation Model
- Google’s next Genie AI model generates more realistic experiences
- Trump Says He'll Ban The Use Of The Term "Artificial Intelligence"
- Trump says US to ban 'artificial intelligence' term, pushes 'super intelligence'
- The Boar
- RTVE launches 'Telerín', a free Artificial Intelligence to discover the secrets of TVE's early years
- Elon Musk says he has ‘cyber psychosis’. Do you too?
- Neuroscience gave AI its start. What happens now?
- AI data centre concerns in Texas ahead of U.S. midterm elections
- Do Humans and AI Even Speak the Same Language?
- I Tested Every Major AI Agent for a Day. Here’s the Best One
- Dell ties supply chain security to agentic AI risks
Comments
Please log in to post a comment.