Privacy-preserving AI gets a boost with Google’s VaultGemma model

Google has unveiled VaultGemma, a new large language model built to offer cutting-edge privacy through differential privacy. The 1-billion-parameter model is based on Google’s Gemma architecture and is described as the most powerful differentially private LLM to date.

Differential privacy adds mathematical noise to data, preventing the identification of individuals while still producing accurate overall results. The method has long been used in regulated industries, but has been challenging to apply to large language models without compromising performance.

VaultGemma is designed to eliminate that trade-off. Google states that the model can be trained and deployed with differential privacy enabled, while maintaining comparable stability and efficiency to non-private LLMs.

This breakthrough could have significant implications for developers building privacy-sensitive AI systems, ranging from healthcare and finance to government services. It demonstrates that sensitive data can be protected without sacrificing speed or accuracy.

Google’s research teams say the model will be released with open-source tools to help others adopt privacy-preserving techniques. The move comes amid rising regulatory and public scrutiny over how AI systems handle personal data.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Scientist and DeepMind’s CEO Demis Hassabis says learning how to learn is key to the AI future

The Nobel laureate, Demis Hassabis, has argued that the most crucial ability for the next generation will be learning how to learn.

Speaking at the Odeon of Herodes Atticus in Athens, Greece, he said adaptability was vital as AI reshapes work and education.

The neuroscientist and former chess prodigy predicted that AGI machines with human-level versatility could emerge within a decade. He described it as a development that may create a future of radical abundance, although he warned of risks.

Hassabis urged a stronger focus on ‘meta-skills’ such as optimising approaches to new subjects, instead of relying solely on traditional disciplines.

Given the speed of technological change, he emphasised that people will need to update their knowledge continuously throughout their careers.

His remarks came during a discussion with Greek Prime Minister Kyriakos Mitsotakis, who warned that the unchecked growth of technology giants could fuel economic inequality and social unrest if citizens do not see clear benefits from AI adoption.

Hassabis’s work on protein folding won him the 2024 Nobel Prize in chemistry.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Google Quantum AI selected for DARPA quantum benchmarking initiative

Google Quantum AI has been selected by the US Defence Advanced Research Projects Agency (DARPA) to participate in the Quantum Benchmarking Initiative (QBI). QBI is designed to assess quantum computing approaches and judge whether utility-scale, fault-tolerant quantum computers could be developed by 2033.

The selection means Google will work with DARPA’s technical experts, who will be independent validators for its quantum computing roadmap. The evaluation aims to provide rigorous third-party benchmarking, a critical capability for the broader quantum industry.

DARPA’s QBI is not only about validation. It aims to compare different quantum technologies, superconducting qubits, photonic systems, trapped ions and other modalities under shared metrics.

Google’s involvement underscores its ongoing mission to build quantum infrastructure capable of addressing problems such as new medicine design, energy innovation and machine-learning optimisation.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

NotebookLM turns notes into flashcards podcasts and quizzes

Google’s learning-focused AI tool NotebookLM has gained a major update, making studying and teaching more interactive.

Instead of offering only static summaries, it now generates flashcards that condense key information into easy-to-remember notes, helping users recall knowledge more effectively.

Reports can also be transformed into quizzes with customisable topics and difficulty, which can then be shared with friends or colleagues through a simple link.

The update extends to audio learning, where NotebookLM’s podcast-style Audio Overviews are evolving with new formats. Instead of a single style, users can now create Brief, Debate, or Critique episodes, giving greater flexibility in how material is explained or discussed.

Google is also strengthening its teaching tools. A new Blog Post format offers contextual suggestions such as strategy papers or explainers, while the ability to create custom report formats allows users to design study resources tailored to their needs.

The most significant addition, however, is the Learning Guide. Acting like a personal tutor, it promotes deeper understanding by asking open-ended questions, breaking problems into smaller steps, and adapting explanations to suit each learner.

With these features, NotebookLM is moving closer to becoming a comprehensive learning assistant, offering a mix of interactive study aids and adaptable teaching methods that go beyond simple note-taking.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

Google launches AI Plus to expand access to Gemini tools

Google has introduced a new mid-tier subscription called AI Plus, designed to make its Gemini AI tools more accessible at a lower price. Positioned between the free and premium Pro plans, AI Plus offers broader access to Gemini 2.5 Pro, productivity features across Gmail, Docs, and Sheets, and 200GB of Google One cloud storage.

Subscribers will benefit from a larger 128,000-token context window, compared with 32,000 for free users, and tools such as Veo 3 Fast for video creation, Google Flow for video generation, and Whisk for image-to-video conversion. The plan also includes expanded use of NotebookLM, Google’s AI-powered research assistant.

The service has launched first in Indonesia at Rp. 75,000 ($4.56) per month, a fraction of the AI Pro plan’s price of Rp. 309,000 ($18.79). Google has not set a timeline for global rollout but said AI Plus will cost under $20 per month in other markets, with prices adjusted regionally.

After months of vague descriptions, the new tier follows Google’s move to publish clearer guidelines on Gemini’s usage limits across free, Pro, and Ultra plans. The update aims to bring greater transparency as the company pushes deeper into the competitive AI subscription market.

Would you like to learn more about AI, tech and digital diplomacyIf so, ask our Diplo chatbot!

Pressure mounts as Apple prepares AI search push with Google ties

Apple’s struggles in the AI race have been hard to miss. Its Apple Intelligence launch was disappointing, and its reliance on ChatGPT appeared to be a concession to rivals.

Bloomberg’s Mark Gurman now reports that Apple plans to introduce its AI-powered web search tool in spring 2026. The move would position it against OpenAI and Perplexity, while renewing pressure on Google.

The speculation comes after news that Google may integrate its Gemini AI into Apple devices. During an antitrust trial in April, Google CEO Sundar Pichai confirmed plans to roll out updates later this year.

According to Gurman, Apple and Google finalised an agreement for Apple to test a Google-developed AI model to boost its voice assistant. The partnership reflects Apple’s mixed strategy of dependence and rivalry with Google.

With a strong record for accurate Apple forecasts, Gurman suggests the company hopes the move will narrow its competitive gap. Whether it can outpace Google, especially given Pixel’s strong AI features, remains an open question.

Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!

Google avoids breakup as AI reshapes search and threatens e-commerce traffic

The US tech giant Google will not be forced to divest Chrome or Android following the long-running US monopoly case.

Judge Mehta ruled that while Google holds a monopoly in traditional search, the rise of AI companies is creating new competitive pressures.

The judgement prevents Google from striking exclusive distribution deals but still allows it to pay partners for preloading and placement of its products. The court also ordered Google to loosen its control over search data, a move that could enable rivals to build their own AI-driven search tools.

Yet, concerns remain for e-commerce businesses.

Google Zero, the company’s AI-powered search summary, is cutting website traffic by keeping users within Google’s results.

Research shows sharp declines in mobile click-through rates, leaving online retailers uncertain of their future visibility.

Experts warn that zero-click searches are becoming the norm. Businesses are being urged to optimise for Google’s AI overviews, enhance the value of product and review pages, track traffic impacts, and diversify their marketing channels.

While Google has avoided structural remedies, its dominance in search and AI appears far from over.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

UK publishers fear Google AI summaries hit revenues

UK publishers warn that Google’s AI Overviews significantly cut website traffic, threatening fragile online revenues.

Reach, owner of the Mirror and Daily Express, said readers often settle for the AI summary instead of visiting its sites. DMG Media told regulators that click-through rates had fallen by up to 89% since the rollout.

Publishers argue that they provide accurate reporting that fuels Google’s search results, yet they see no financial return when users no longer click through. Concerns are growing over Google’s conversational AI Mode, which displays even fewer links.

Google insists that search traffic has remained stable year-on-year and claims that AI overviews offer users more opportunities to find quality links. Still, a coalition of publishers has filed a complaint with the UK Competition and Markets Authority, alleging misuse of their content.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot

Google boosts Gemini with audio uploads and NotebookLM upgrades

The US tech giant has expanded the capability of its Gemini app by allowing users to upload audio files for AI analysis across Android, iOS, and the web. The upgrade enables transcription of interviews, voice memos and lecture recordings instead of relying solely on typed or spoken prompts.

Free-tier users can upload clips of up to ten minutes with five prompts daily, while paid subscribers have access to three hours of uploads across multiple files. According to Gemini vice president Josh Woodward, the feature is designed to make the platform more versatile and practical for everyday tasks.

Google has also enhanced its Search AI mode with five new languages, including Hindi, Japanese and Korean, extending its multilingual reach.

NotebookLM, the company’s research assistant powered by Gemini, can now generate structured reports such as quizzes, study guides and blog posts from uploaded content, available in more than 80 languages.

These improvements underline Google’s ambition to integrate AI more deeply into everyday applications instead of leaving the technology confined to experimental tools. They also highlight growing competition in the AI market, with Google using Gemini 2.5 to expand its services for global users.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!

AI Mode in Google Search adds multilingual support to Hindi and four more languages

Google has announced an expansion of AI Mode in Search to five new languages, including Hindi, Indonesian, Japanese, Korean and Brazilian Portuguese. The feature was first introduced in English in March and aims to compete with AI-powered search platforms such as ChatGPT Search and Perplexity AI.

The company highlighted that building a global search experience requires more than translation. Google’s custom version of Gemini 2.5 uses advanced reasoning and multimodal capabilities to provide locally relevant and useful search results instead of offering generic answers.

AI Mode now also supports agentic tasks such as booking restaurant reservations, with plans to include local service appointments and event ticketing.

Currently, these advanced functions are available to Google AI Ultra subscribers in the US, while India received the rollout of the language expansion in July.

These developments reinforce Google’s strategy to integrate AI deeply into its search ecosystem, enhancing user experience across diverse regions instead of limiting sophisticated AI tools to English-language users.

Would you like to learn more about AI, tech and digital diplomacy? If so, ask our Diplo chatbot!