Meta has introduced new AI-powered translation features that allow Facebook and Instagram users to enjoy reels from around the world in multiple languages.
Meta AI now translates, dubs, and lip-syncs short videos in English, Spanish, Hindi, and Portuguese, with more languages to be added soon.
A tool that reproduces a creator’s voice and tone while automatically syncing translated audio to their lip movements, providing a natural viewing experience. It is free for Facebook creators with over 1,000 followers and all public Instagram accounts in countries where Meta AI is available.
The expansion is part of Meta’s goal to make global content more accessible and to help creators reach wider audiences. By breaking language barriers, Meta aims to strengthen community connections and turn Reels into a platform for global cultural exchange.
Would you like to learn more aboutAI, tech and digital diplomacy? If so, ask our Diplo chatbot!
California has entered a new era of privacy and AI enforcement after the state’s privacy regulator fined Tractor Supply USD1.35 million for failing to honour opt-outs and ignoring Global Privacy Control signals. The case marks the largest penalty yet from the California Privacy Protection Agency.
In California, there is a widening focus on how companies manage consumer data, verification processes and third-party vendors. Regulators are now demanding that privacy signals be enforced at the technology layer, not just displayed through website banners or webforms.
Retailers must now show active, auditable compliance, with clear privacy notices, automated data controls and stronger vendor agreements. Regulators have also warned that businesses will be held responsible for partner failures and poor oversight of cookies and tracking tools.
At the same time, California’s new AI law, SB 53, extends governance obligations to frontier AI developers, requiring transparency around safety benchmarks and misuse prevention. The measure connects AI accountability to broader data governance, reinforcing that privacy and AI oversight are now inseparable.
Executives across retail and technology are being urged to embed compliance and governance into daily operations. California’s regulators are shifting from punishing visible lapses to demanding continuous, verifiable proof of compliance across both data and AI systems.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Ant Group has unveiled its Ling AI model family, introducing Ling-1T, a trillion-parameter large language model that has been open-sourced for public use.
The Ling family now includes three main series: the Ling non-thinking models, the Ring thinking models, and the multimodal Ming models.
Ling-1T delivers state-of-the-art performance in code generation, mathematical reasoning, and logical problem-solving, achieving 70.42% accuracy on the 2025 AIME benchmark.
A model that combines efficient inference with strong reasoning capabilities, marking a major advance in AI development for complex cognitive tasks.
Company’s Chief Technology Officer, He Zhengyu, said that Ant Group views AGI as a public good that should benefit society.
The release of Ling-1T and the earlier Ring-1T-preview underscores Ant Group’s commitment to open, collaborative AI innovation and the development of inclusive AGI technologies.
Would you like to learn more aboutAI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Researchers at MIT’s Computer Science and AI Laboratory have developed a new AI system that can build realistic virtual environments for training robots. The tool, called steerable scene generation, creates kitchens, restaurants and living rooms filled with 3D objects where robots interact with the physical world.
The system uses a diffusion model guided by Monte Carlo tree search to produce scenes that follow real-world physics. Unlike traditional simulations, it can accurately position objects and avoid visual errors such as items overlapping or floating unrealistically.
By generating millions of unique, lifelike environments, the system can dramatically increase the training data available for robotic foundation models. Robots trained in these AI settings can practise everyday actions like stacking plates or placing cutlery with greater precision.
The researchers say the technique allows robots to learn more efficiently without the cost or limits of real-world testing. Future work aims to include movable objects and internet-sourced assets to make the simulations even more dynamic and diverse.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Microsoft Azure has launched the world’s first NVIDIA GB300 NVL72 supercomputing cluster, explicitly designed for OpenAI’s large-scale AI workloads.
The new NDv6 GB300 VM series integrates over 4,600 NVIDIA Blackwell Ultra GPUs, representing a significant step forward in US AI infrastructure and innovation leadership.
Each rack-scale system combines 72 GPUs and 36 Grace CPUs, offering 37 terabytes of fast memory and 1.44 exaflops of FP4 performance.
A configuration that supports complex reasoning and multimodal AI systems, achieving up to five times the throughput of the previous NVIDIA Hopper architecture in MLPerf benchmarks.
The cluster is built on NVIDIA’s Quantum-X800 InfiniBand network, delivering 800 Gb/s of bandwidth per GPU for unified, high-speed performance.
Microsoft and NVIDIA’s long-standing collaboration has enabled a system capable of powering trillion-parameter models, positioning Azure at the forefront of the next generation of AI training and deployment.
Would you like to learn more aboutAI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Under the deal, Microsoft will pay a licensing fee to Harvard. The partnership aims to enhance Microsoft’s Copilot AI assistant by enabling it to provide medical advice that more closely aligns with what a user might receive from a healthcare professional.
So far, Copilot’s underlying models have been powered primarily by OpenAI. But this agreement is part of Microsoft’s broader push to diversify and reduce its dependence on OpenAI’s technology stack.
Dominic King, Microsoft’s vice president of health, has said the goal is for Copilot’s responses to health queries to more accurately reflect what a medical practitioner would say, rather than generic or superficial answers.
Microsoft declined to comment in detail, but the move strengthens Copilot’s differentiation in the health domain, arguably a high-stakes vertical area where accuracy and trustworthiness matter greatly.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
The Nobel Prize in Physics has spotlighted quantum mechanics’ growing role in shaping a smarter, more sustainable future. Such advances are reshaping technology across communications and energy.
Researchers are finding new ways to use quantum effects to boost efficiency. Quantum computing could ease AI’s power demands, while novel production methods may transform energy systems.
A Institute of Science Tokyo team has built a quantum energy harvester that captures waste heat and converts it into power, bypassing traditional thermodynamic limits.
MIT has observed frictionless electron movement, and new quantum batteries promise faster charging by storing energy in photons. The breakthroughs could enable cleaner and more efficient technologies.
Quantum advances offer huge opportunities but also risks, including threats to encryption. Responsible governance will be crucial to ensure these technologies serve the public good.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!
Nvidia is reportedly investing up to $2bn in Elon Musk’s AI company, xAI, as part of a $20bn funding round aimed at scaling its Colossus 2 data centre in Memphis. The capital will be used to buy Nvidia GPUs, essential for powering xAI’s next generation of AI models.
The funding package combines about $7.5bn in equity and up to $12.5bn in debt, structured through a special purpose vehicle that will lease the hardware to xAI over five years. The debt is secured by the GPUs themselves, allowing investors to recover their costs through chip rentals.
xAI faces mounting financial pressure, with reports indicating a cash burn of around $1bn per month. The firm raised $10bn earlier in the year and continues to draw on capital from Musk’s other ventures, including SpaceX.
The move comes amid an intense funding surge across the AI sector, as OpenAI, Meta and Oracle also announce multi-billion-dollar investments in infrastructure. Nvidia’s latest deal with xAI further cements its position at the centre of the global AI hardware ecosystem.
Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!
Researchers are using satellite imagery and AI modelling to map global mining activity and close critical data gaps. Transition minerals, such as lithium and copper, are vital for renewable technologies but often come from ecologically sensitive regions, raising concerns about both environmental and social impacts.
Project lead Victor Maus from the Vienna University of Economics and Business said many new projects overlap with areas of high biodiversity or Indigenous lands. Over half of transition mineral resources are on or near Indigenous or subsistence farming territories, according to earlier studies.
Previous mapping efforts have struggled to document small-scale and informal mining, which remains unregulated despite its impact. Maus’s team compared satellite images of 120,000 square kilometres of mine footprints with the S&P Capital IQ Pro database and found over half missing.
To close these gaps, the team is creating a mining database under the EU-funded Mine the Gap initiative. By combining multispectral, radar, and hyperspectral imagery with AI, they aim to monitor land use, waste generation, and environmental degradation.
Experts say the database could support policymakers and increase transparency. Maus emphasised that global reporting standards are crucial for enhancing accountability and informing decisions on managing the environmental and social impacts of mining.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!
Discord has confirmed that official ID images belonging to around 70,000 users may have been exposed in a cyberattack targeting a third-party service provider. The platform itself was not breached, but hackers targeted a company involved in age verification processes.
The leaked data may include personal information, partial credit card details, and conversations with Discord’s customer service agents. No full credit card numbers, passwords, or activity beyond support interactions were affected. Impacted users have been contacted, and law enforcement is investigating.
The platform has revoked the support provider’s access to its systems and has not named the third party involved. Zendesk, a customer service software supplier to Discord, said its own systems were not compromised and denied being the source of the breach.
Discord has rejected claims circulating online that the breach was larger than reported, calling them part of an attempted extortion. The company stated it would not comply with demands from the attackers. Cybercriminals often sell personal information on illicit markets for use in scams.
ID numbers and official documents are especially valuable because, unlike credit card details, they rarely change. Discord previously tightened its age-verification measures following concerns over the misuse of some servers to distribute illegal material.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!
