Google

Gemini Robotics On-Device: Google’s AI model for offline robotic tasks

On Tuesday, 24 June, Google’s DeepMind division announced the release of a new large language model named Gemini Robotics On-Device, designed to operate locally on robotic systems.

In a blog post, the company stated that the AI model has been optimised to function efficiently on-device and demonstrates strong general-purpose dexterity and task generalisation capabilities.

The offline model is an advancement of the earlier Gemini Robotics system introduced in March this year. Unlike cloud-based models, this version can operate offline, making it suitable for limited connectivity or critical latency.

Engineered for robots with dual arms, Gemini Robotics On-Device is designed to require minimal computational resources.

It can execute fine motor tasks such as folding garments and unzipping bags. According to Google, the model responds to natural language prompts, enabling more intuitive human-robot interaction.

The company claims the model outperforms comparable on-device alternatives, especially when completing complex, multi-step instructions or handling unfamiliar tasks. Benchmark results indicate that its performance closely approaches that of Google’s cloud-based AI solutions.

Initially developed for ALOHA robots, the on-device model has since been adapted for other systems, including the bi-arm Franka FR3 robot and the Apollo humanoid.

On the Franka FR3, the model followed diverse instructions and managed unfamiliar objects and environments, including industrial tasks like belt assembly. The system demonstrated general object manipulation in previously unseen contexts on the Apollo humanoid.

Developers interested in trialling Gemini Robotics On-Device can access it via the provided software development kit (SDK).

Google joins other major players exploring AI for robotics. At GTC 2025, NVIDIA introduced Groot N1, an AI system for humanoid robots, while Hugging Face is currently developing its own open-source, AI-powered robotic platform.

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

Google launches AI Mode Search in India

Google has launched its advanced AI Mode search experience in India, allowing users to explore information through more natural and complex interactions.

The feature, previously available as an experiment in the US, can now be enabled in English via Search Labs. Users test experimental tools on this platform and share feedback on early Google Search features.

Once activated, AI Mode introduces a new tab in the Search interface and Google app. It offers expanded reasoning capabilities powered by Gemini 2.5, enabling queries through text, voice, or images.

The shift supports deeper exploration by allowing follow-up questions and offering diverse web links, helping users understand topics from multiple viewpoints.

India plays a key role in this rollout due to its widespread visual and voice search use.

According to Hema Budaraju, Vice President of Product Management for Search, more users in India engage with Google Lens each month than anywhere else. AI Mode reflects Google’s broader goal of making information accessible across different formats.

Google also highlighted that over 1.5 billion people globally use AI Overviews monthly. These AI-generated summaries, which appear at the top of search results, have driven a 10% rise in user engagement for specific types of queries in both India and the US.

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

Apple considers buying Perplexity AI

Apple is reportedly considering the acquisition of Perplexity AI as it attempts to catch up in the fast-moving race for dominance in generative technology.

According to Bloomberg, the discussions involve senior executives, including Eddy Cue and merger head Adrian Perica, who remain at an early stage.

Such a move would significantly shift Apple, which typically avoids large-scale takeovers. However, with investor pressure mounting after an underwhelming developer conference, the tech giant may rethink its traditionally cautious acquisition strategy.

Perplexity has gained prominence for its fast, clear AI chatbot and recently secured funding at a $14 billion valuation.

Should Apple proceed, the acquisition would be the company’s largest ever financially and strategically, potentially transforming its position in AI and reducing its long-standing dependence on Google’s search infrastructure.

Apple’s slow development of Siri and reliance on a $20 billion revenue-sharing deal with Google have left it trailing rivals. With that partnership now under regulatory scrutiny in the US, Apple may view Perplexity as a vital step towards building a more autonomous search and AI ecosystem.

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

EU adviser backs Android antitrust ruling against Google

An adviser to the Court of Justice of the European Union has supported the EU’s antitrust ruling against Google, recommending the dismissal of its appeal over a €4.1bn fine. The case concerns Google’s use of its Android mobile system to limit competition through pre-installed apps and contractual restrictions.

The original €4.34bn fine was imposed by the European Commission in 2018 and later reduced by the General Court.

Google then appealed to the EU’s top court, but Advocate-General Juliane Kokott concluded that Google’s practices gave it unfair market advantages.

Kokott rejected Google’s argument that its actions should be assessed against an equally efficient competitor, noting Google’s dominance in the Android ecosystem and the robust network effects it enjoys.

She argued that bundling Google Search and Chrome with the Play Store created barriers for competitors.

The final court ruling is expected in the coming months and could shape Google’s future regulatory obligations in Europe. Google has already incurred over €8 billion in the EU antitrust fines across several investigations.

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

Massive data leak exposes 16 billion login credentials from Google, Facebook, and more

One of the largest-ever leaks of stolen login data has come to light, exposing more than 16 billion records across widely used services, including Facebook, Google, Telegram, and GitHub. The breach, uncovered by researchers at Cybernews, highlights a growing threat to individuals and organisations.

The exposed data reportedly originated from info stealer malware, previous leaks, and credential-stuffing tools. A total of 30 separate datasets were identified, some containing over 3.5 billion entries.

These were briefly available online due to unsecured cloud storage before being removed. Despite the swift takedown, the data had already been collected and analysed.

Experts have warned that the breach could lead to identity theft, phishing, and account takeovers. Smaller websites and users with poor cybersecurity practices are especially vulnerable. Many users continue to reuse passwords or minor variations of them, increasing the risk of exploitation.

While the leak is severe, users employing two-factor authentication (2FA), password managers, or passkeys are less likely to be affected.

Passkeys, increasingly adopted by companies like Google and Apple, offer a phishing-resistant login method that bypasses the need for passwords altogether.

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

Google launches AI voice chat in Search app for Android and iOS

Google has started rolling out its new ‘Search Live in AI Mode’ for the Google app on Android and iOS, offering users the ability to have seamless voice-based conversations with Search.

Currently available only in the US for those signed up to the AI Mode experiment in Labs, the feature was previewed at last month’s Google I/O conference.

The tool uses a specially adapted version of Google’s Gemini AI model, fine-tuned to deliver smarter voice interactions. It combines the model’s capabilities with Google Search’s information infrastructure to provide real-time spoken responses.

Using a technique called ‘query fan-out’, the system retrieves a wide range of web content, helping users discover more varied and relevant information.

The new mode is particularly useful when multitasking or on the go. Users can tap a ‘Live’ icon in the Google app and ask spoken queries like how to keep clothes from wrinkling in a suitcase.

Follow-up questions are handled just as naturally, and related links are displayed on-screen, letting users read more without breaking their flow.

To use the feature, users can tap a sparkle-shaped waveform icon under the Search bar or next to the search field. Once activated, a full-screen interface appears with voice control options and a scrolling list of relevant links.

Even with the phone locked or other apps open, the feature keeps running. A mute button, transcript view, and voice style settings—named Cassini, Cosmo, Neso, and Terra—offer additional control over the experience.

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

AI helps Google curb scams and deepfakes in India

Google has introduced its Safety Charter for India to combat rising online fraud, deepfakes and cybersecurity threats. The charter outlines a collaborative plan focused on user safety, responsible AI development and protection of digital infrastructure.

AI-powered measures have already helped Google detect 20 times more scam-related pages, block over 500 million scam messages monthly, and issue 2.5 billion suspicious link warnings. Its ‘Digikavach’ programme has reached over 177 million Indians with fraud prevention tools and awareness campaigns.

Google Pay alone averted financial fraud worth ₹13,000 crore in 2024, while Google Play Protect stopped nearly 6 crore high-risk app installations. These achievements reflect the company’s ‘AI-first, secure-by-design’ strategy for early threat detection and response.

The tech giant is also collaborating with IIT-Madras on post-quantum cryptography and privacy-first technologies. Through language models like Gemini and watermarking initiatives such as SynthID, Google aims to build trust and inclusion across India’s digital ecosystem.

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

Google warns against weak passwords amid £12bn scams

Gmail users are being urged to upgrade their security as online scams continue to rise sharply, with cyber criminals stealing over £12 billion in the past year alone. Google is warning that simple passwords leave people vulnerable to phishing and account takeovers.

To combat the threat, users are encouraged to switch to passkeys or use ‘Sign in with Google’, both of which offer stronger protections through fingerprint, face ID or PIN verification. Over 60% of Baby Boomers and Gen X users still rely on weak passwords, increasing their exposure to attacks.

Despite the availability of secure alternatives, only 30% of users reportedly use them daily. Gen Z is leading the shift by adopting newer tools, bypassing outdated security habits altogether.

Google recommends adding 2-Step Verification for those unwilling to leave passwords behind. With scams growing more sophisticated, extra security measures are no longer optional, they are essential.

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

Quantum computing threatens Bitcoin: Experts debate timeline

Recent breakthroughs in quantum computing have revived fears about the long-term security of Bitcoin (BTC).

With IBM aiming to release the first fault-tolerant quantum computer, the IBM Quantum Starling, by 2029, experts are increasingly concerned that such advancements could undermine Bitcoin’s cryptographic backbone.

Bitcoin currently relies on elliptic curve cryptography (ECC) and the SHA-256 hashing algorithm to secure wallets and transactions. However, both are potentially vulnerable to Shor’s algorithm, which a sufficiently powerful quantum computer could exploit.

Google quantum researcher Craig Gidney warned in May 2025 that quantum resources required to break RSA encryption had been significantly overestimated. Although Bitcoin uses ECC, not RSA, Gidney’s research hinted at a threat window between 2030 and 2035 for crypto systems.

Opinions on the timeline vary. Adam Back, Blockstream CEO and early Bitcoin advocate, believes a quantum threat is still at least two decades away. However, he admitted that future progress could force users to migrate coins to quantum-safe wallets—potentially even Satoshi Nakamoto’s dormant holdings.

Others are more alarmed. David Carvalho, CEO of Naoris Protocol, claimed in a June 2025 op-ed that Bitcoin could be cracked within five years, pointing to emerging technologies like Microsoft’s Majorana chip. He estimated that nearly 30% of BTC is stored in quantum-vulnerable addresses.

‘Just one breach could destroy trust in the entire ecosystem,’ Carvalho warned, noting that BlackRock has already acknowledged the quantum risk in its Bitcoin ETF filings.

Echoing this urgency, billionaire investor Chamath Palihapitiya said in late 2024 that SHA-256 could be broken within two to five years if companies scale quantum chips like Google’s 105-qubit Willow. He urged the crypto industry to start updating encryption protocols before it’s too late.

While truly fault-tolerant quantum machines capable of breaking Bitcoin are not yet available, the accelerating pace of research suggests that preparing for a quantum future is no longer optional—it’s a necessity.

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

OpenAI turns to Google Cloud in shift from solo AI race

OpenAI has entered into an unexpected partnership with Google, using Google Cloud to support its growing AI infrastructure needs.

Despite being fierce competitors in AI, the two tech giants recognise that long-term success may require collaboration instead of isolation.

As the demand for high-performance hardware soars, traditional rivals join forces to keep pace. OpenAI, previously backed heavily by Microsoft, now draws from Google’s vast cloud resources, hinting at a changing attitude in the AI race.

Rather than going it alone, firms may benefit more by leveraging each other’s strengths to accelerate development.

Google CEO Sundar Pichai, speaking on a podcast, suggested there is room for multiple winners in the AI sector. He even noted that a major competitor had ‘invited me to a dance’, underscoring a new phase of pragmatic cooperation.

While Google still faces threats to its search dominance from tools like ChatGPT, business incentives may override rivalry.

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