Artificial intelligence

Court ruling raises alarm over saved ChatGPT chats

A US federal court has ordered OpenAI to preserve nearly all user chats with ChatGPT, including those that users had deleted. The decision comes as part of The New York Times’s ongoing copyright lawsuit, triggering widespread privacy concerns.

The ruling means that millions of personal conversations, previously thought erased, will remain accessible during litigation. These exchanges may include medical queries, relationship issues, and other private matters shared in confidence.

Privacy advocates argue that users were not notified or allowed to object. Critics warn the US ruling sets a dangerous precedent, enabling mass data preservation in lawsuits unrelated to most users.

The Times claims users may have deleted chats to hide copyright infringement. Lawyers and privacy experts counter that people delete chats for legitimate, non-infringing reasons and should retain control over their data.

Legal experts call the preservation order excessive, noting it undermines trust in AI tools and could lead to a chilling effect on their use. The decision could reshape how user privacy is treated in tech litigation for years.

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

Podcast-based training helps improve AI dialogue

Researchers have developed PodGPT, a new AI model designed to enhance reasoning and dialogue skills by training on scientific podcasts. The project aims to integrate dynamic, conversational audio data into language models to boost their performance in STEMM subjects.

The team used over 3,700 hours of English-language STEMM podcast transcripts, alongside material from the New England Journal of Medicine. Transcripts were generated using Whisper large-v3 and fed into open-source AI models such as Gemma, Mixtral, and LLaMA.

PodGPT improves multilingual understanding and factual accuracy, particularly in answering science-based queries. It also performs better at retrieving evidence from long documents and engaging in human-like scientific dialogue.

The researchers suggest that podcast-based training provides more realistic language use and diverse reasoning patterns than traditional datasets. Their work demonstrates the value of spoken, expert-led content in preparing models for advanced scientific applications.

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

Google hit with EU complaint over AI Overviews

After a formal filing by the Independent Publishers Alliance, Google has faced an antitrust complaint in the European Union over its AI Overviews feature.

The group alleges that Google has been using web content without proper consent to power its AI-generated summaries, causing considerable harm to online publishers.

The complaint claims that publishers have lost traffic, readers and advertising revenue due to these summaries. It also argues that opting out of AI Overviews is not a real choice unless publishers are prepared to vanish entirely from Google’s search results.

AI Overviews were launched over a year ago and now appear at the top of many search queries, summarising information using AI. Although the tool has expanded rapidly, critics argue it drives users away from original publisher websites, especially news outlets.

Google has responded by stating its AI search tools allow users to ask more complex questions and help businesses and creators get discovered. The tech giant also insisted that web traffic patterns are influenced by many factors and warned against conclusions based on limited data.

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

EU rejects delay for AI Act rollout

The EU has confirmed it will enforce its originally scheduled AI Act, despite growing calls from American and European tech firms to delay the rollout.

Major companies, including Alphabet, Meta, ASML and Mistral, have urged the European Commission to push back the timeline by several years, citing concerns over compliance costs.

Rejecting the pressure, a Commission spokesperson clarified there would be no pause or grace period. The legislation’s deadlines remain, with general-purpose AI rules taking effect this August and stricter requirements for high-risk systems following August 2026.

The AI Act represents the EU’s effort to regulate AI across various sectors, aiming to balance innovation and public safety. While tech giants argue that the rules are too demanding, the EU insists legal certainty is vital and the framework must move forward as planned.

The Commission intends to simplify the process later in the year, such as easing reporting demands for smaller businesses. Yet the core structure and deadlines of the AI Act will not be altered.

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

BRICS calls for AI data regulations amid challenges with de-dollarisation

BRICS leaders in Rio de Janeiro have called for stricter global rules on how AI uses data, demanding fair compensation for content used without permission.

The group’s draft statement highlights growing frustration with tech giants using vast amounts of unlicensed content to train AI models.

Despite making progress on digital policy, BRICS once again stalled on a long-standing ambition to reduce reliance on the US dollar.

After a decade of talks, the bloc’s cross-border payments system remains in limbo. Member nations continue to debate infrastructure, governance and how to work around non-convertible currencies and sanctions.

China is moving independently, expanding the yuan’s international use and launching domestic currency futures.

Meanwhile, the rest of the bloc struggles with legal, financial and technical hurdles, leaving the dream of a unified alternative to the dollar on hold. Even a proposed New Investment Platform remains mired in internal disagreements.

In response to rising global debt concerns, BRICS introduced a Multilateral Guarantees Initiative within the New Development Bank. It aims to improve credit access across the Global South without needing new capital, especially for countries struggling to borrow in dollar-dominated markets.

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

Ooredoo launches local AI cloud service in Qatar

Ooredoo has unveiled a new sovereign AI cloud service powered by NVIDIA Hopper GPUs, hosted locally in Qatar’s data centres. The move supports the country’s broader aim to advance digital transformation and position itself as a regional leader in AI innovation.

The service accelerates AI adoption across key sectors such as energy, finance, logistics, healthcare and smart cities. As an NVIDIA Cloud Partner, Ooredoo now provides customers access to cutting-edge GPU technology and the NVIDIA AI Enterprise platform for developing and scaling AI solutions.

Use cases include real-time financial analysis, supply chain optimisation and chatbot development, all supported by high-performance cloud computing infrastructure operated by Syntys. These capabilities aim to deliver robust local AI services that meet national security and sovereignty requirements.

The initiative aligns with the Qatar Digital Agenda 2030 and the National AI Strategy, which advocate for local hosting, advanced digital infrastructure and responsible AI development. Ooredoo’s CEO, Sheikh Ali Bin Jabor Al-Thani, said the project will drive economic growth and innovation across all levels of society.

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

India needs a quantum leap in defence AI, says LatentAI founder

Jags Kandasamy, founder of US-based defence tech company LatentAI, is working with Indian firms to pursue defence contracts, but says India must leapfrog forward in AI-enabled warfare. In an interview with HT, he outlined the challenges and opportunities in the India-US defence tech partnership.

At Aero India, Kandasamy saw an Indian Army prototype using computer vision on automated border weapons. While promising, the system’s heavy computing requirements limit scalability.

LatentAI, which helped the US Navy optimise AI models for underwater vehicles, offers solutions by compressing models to fit limited hardware. ‘Autonomous warfare is the future,’ he said, noting the impact of drones and AI on intelligence, targeting and surveillance.

Kandasamy’s India partner, InferQ, came through introductions by Forge Ventures, which works closely with Indian and United States defence departments. He said that government initiatives like INDUSX and IDEX are helping firms connect across borders, but procedural bottlenecks persist.

‘There’s no mechanism for non-Indian passport holders to get clearance,’ Kandasamy noted. ‘In the US, the process is transparent. Even Indian firms can’t see who makes the decisions.’

He recalled advising a founder who waited two years to get a prototype certified. ‘India can be a great proving ground, but the bureaucracy needs streamlining.’

On China, Kandasamy didn’t mince words: ‘China is a third-year PhD student in defence AI. India is in elementary school.’ He cited examples of China’s proactive investment in emerging tech from 2010, including personal offers to relocate his startups. ‘India needs to leapfrog like it did with telecom.’

Still, Kandasamy praised India’s IDEX programme for uncovering strong homegrown tech talent. ‘There are smart people and ideas, but they need support and scale.’

To improve bilateral cooperation, he suggested reciprocal security clearances and defining interoperability frameworks between US and Russian-origin systems in India. ‘India won’t abandon its Russian hardware. But if we can protect both sides’ secrets while working together, that would be real progress.’

‘India and the US are linchpins of a democratic society,’ he concluded. ‘We have to make this partnership work.’

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

Vietnam adopts AI in elite sports training

Vietnam is integrating AI technology into its national sports training programmes to improve performance in international competitions. A new partnership between the Sports Authority of Việt Nam (SAV) and Dreamax Company will see AI applied from 2025 to 2030.

The Vietnamese technology will analyse performance data, monitor adaptation to training, and predict injury risk using devices that collect environmental and spatial indicators. The first implementations are set to begin next week, marking a shift towards data-driven performance evaluation and coaching.

SAV Director Nguyễn Danh Hoàng Việt said the initiative addresses outdated systems and builds on the country’s recent sporting successes by offering more sophisticated and personalised training. Dreamax stated that the AI system supports national management and provides public insight into sporting progress.

Collected data, managed as national security information, will be analysed to measure the programme’s impact over its first year. If successful, the AI-based approach may expand to other teams as Vietnam targets greater achievements in future Asian Games and Olympic events.

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

AI tool could change marine forecasting methods

An AI-driven forecasting tool developed by the Met Office and the University of Exeter is poised to reshape how marine operations are planned. The low-cost model, MaLCOM, has successfully predicted ocean currents in the Gulf of Mexico.

Designed initially to forecast regional wave patterns around the UK, the framework’s adaptability is now helping model ocean currents in new environments.

The tool’s ability to run on a laptop makes it highly accessible, offering real-time insights that could aid offshore energy.

Researchers emphasise the importance of the model’s transparency, which allows users to inspect how it processes data and generates forecasts. This design supports trust in its outputs and offers a strong foundation for ongoing development.

The project began five years ago and has grown through collaboration between academia, government and industry.

Its recent recognition with the ASCE Offshore Technology Conference Best Paper Award underscores the value of partnerships in accelerating progress in AI-based weather and climate tools.

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

Council of Europe picks Jylo to power AI platform

The Council of Europe has chosen Jylo, a European enterprise AI provider, to support over 3,000 users across its organisation.

The decision followed a competitive selection process involving multiple AI vendors, with Jylo standing out for its regulatory compliance and platform adaptability.

As Europe’s leading human rights body, the Council aims to use AI responsibly to support its legal and policy work. Jylo’s platform will streamline document-based workflows and reduce administrative burdens, helping staff focus on critical democratic and legal missions.

Leaders from both Jylo and the Council praised the collaboration. Jylo CEO Shawn Curran said the partnership reflects shared values around regulatory compliance and innovation.

The Council’s CIO, John Hunter, described Jylo’s commitment to secure AI as a perfect fit for the institution’s evolving digital strategy.

Jylo’s AI Assistant and automation features are designed specifically for knowledge-driven organisations. The rollout is expected to strengthen the Council’s internal efficiency and reinforce Jylo’s standing as a trusted AI partner across the European public and legal sectors.

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