To help you stay on top of the latest news, our AI practice group has compiled a roundup of the developments we are following.

• Google’s DeepMind AI subsidiary has revealed a new set of AI tools built upon Gemini 2.0, the latest version of its flagship large language model, according to the MIT Technology Review. DeepMind claims that Gemini 2.0 is twice as fast as the previous version, Gemini 1.5, and directly competes with the latest offerings from rivals OpenAI, Anthropic and Meta. The leading tool unveiled is Project Astra, a multimodal “universal assistant” that can interact through text, speech, images and video, seamlessly integrating with existing Google apps like Search, Maps and Lens. During a closed-door demo, Astra demonstrated capabilities like reading cookbook recipes, recommending wine pairings, identifying artworks, and maintaining context across conversations, though the technology is still in early stages. Alongside Astra, Google introduced complementary AI tools like Mariner (a web-browsing agent), Jules (a coding assistant), and a Gemini-powered assistant for video game tips, signaling the company’s ambitious push to turn generative AI into a mass-market technology. DeepMind also recently unveiled Genie 2, a “foundation world model” that is capable of generating controllable 3D environments from a text or image prompt, procedurally generating frame after frame and responding to user input. DeepMind intends to deploy the technology to train AI agents and aid future AI research.
• A group of U.S. content publishers and creators have warned the United Kingdom’s government over proposed rules for AI data mining, according to the Financial Times. The rules would allow companies to freely scrape web content to train AI models unless the owners of such content affirmatively opt out. In a three-page letter, the Copyright Alliance—representing major American media companies like Disney, Fox, Paramount, Universal Music and Getty—strongly opposed these potential changes, arguing that introducing AI exceptions to copyright laws would discourage creators and rights holders from investing in creative endeavors in the U.K. The letter emphasized that copyright forms the backbone of U.S.-U.K. creative collaborations, citing iconic franchises like Harry Potter and James Bond, and warned that weakening copyright protections could have a chilling effect on future cross-border creative investments. The U.K. government has indicated that its upcoming consultation on AI and creative industries will be “genuine,” and that “no decision has been made.”
• Nvidia has increased its staff in China over the past year to focus on several projects in the country, including AI-powered autonomous vehicles, according to Bloomberg. The company has grown its Chinese workforce from about 3,000 to 4,000 employees in 2024, with approximately 200 new hires in Beijing specifically dedicated to self-driving technology research. Despite facing U.S. trade restrictions that limit its ability to sell advanced semiconductors in China, Nvidia continues to view the country as a crucial market and research hub, generating $5.4 billion in sales in the September quarter. The company’s expansion comes amid a complex geopolitical backdrop, including a recent antitrust probe into Nvidia’s 2020 Mellanox acquisition and ongoing U.S.-China trade tensions, especially with the looming additional tariffs expected when President-elect Donald Trump takes office. Nvidia’s research efforts in China aim to collaborate with local electric vehicle makers and leverage the country’s leadership in EV technologies, with the company hoping to advance its work in driving automation and AI software.
• Startup companies in India are developing AI-powered tools to support agriculture in the country, according to IEEE Spectrum. These tools are aimed at addressing challenges faced by small farmers, including water scarcity, climate unpredictability, and inefficient supply chains. Companies like Fasal and Cropin are providing solutions such as IoT-based weather and soil monitoring, predictive analytics, and AI satellite imagery analysis to offer precise and personalized farming recommendations. The Indian government is also heavily investing in “agritech,” as it recently announced nearly $300 million to support digital agriculture projects. The government is also supporting initiatives such as Agri Stack, a national database of farmers, farm plots, and soil data that will be accessible to the government and agritech companies that is expected to be launched early next year. While these technologies show promise in improving productivity and sustainability, challenges such as affordability, trust and equitable access remain critical concerns.
• Apple is working with Broadcom to develop its own AI processor chips as it seeks to avoid reliance on market leaders such as Nvidia, according to Reuters. The chip, internally code-named Baltra, is expected to be ready for mass production by 2026 and will be manufactured using Taiwan Semiconductor Manufacturing Co’s (TSMC) advanced N3P process. This move aligns with Apple’s recent success in developing in-house chips, such as its M-series processors for Mac laptops, and follows the company’s June announcement of plans to use server chips to power AI features on its devices. The collaboration comes after Apple and Broadcom’s previous multi-billion-dollar partnership for 5G radio frequency components, and reflects a broader trend among big tech companies seeking to diversify their semiconductor supply chains. Broadcom has been a significant beneficiary of the generative AI boom, with its shares gaining 54% in 2024, and the total market for custom chips potentially growing to $45 billion by 2028.

[View source.]