Sustainable development Archives | Page 184 of 304

DeepSeek struggles to launch R2 amid NVIDIA chip shortage

The launch of DeepSeek’s next-generation AI model, R2, is expected to face delays due to a shortage of NVIDIA H20 chips in China.

These chips, designed specifically for the Chinese market following US export restrictions, are essential for running DeepSeek’s highly optimised models.

The ban on H20 shipments in April has triggered widespread concern among cloud providers about the scalability of R2, especially if it outperforms existing open-source models.

CEO Liang Wenfeng has reportedly held back the model’s release, expressing dissatisfaction with its current performance.

Engineers continue refining R2, but the lack of compatible hardware poses a deeper challenge. DeepSeek’s reliance on NVIDIA architecture makes switching to Chinese chips inefficient, as the models are tightly built for NVIDIA’s software and hardware ecosystem.

Some Chinese firms have begun using workarounds by flying engineers to Malaysia, where NVIDIA chips are still available in local data centres.

After training their models abroad, teams return to China with trained systems. Others rely on gaming GPUs like the RTX 5090, which are easier to access via grey markets despite restrictions.

While Chinese tech giants ordered 1.2 million H20 chips earlier in 2025 to meet demand sparked by R1’s success, inventory is still unlikely to support a full R2 rollout.

Companies outside China may launch R2 more easily without facing the same export hurdles.

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

Meta hires top OpenAI researcher for AI superintelligence push

Meta has reportedly hired AI researcher Trapit Bansal, who previously worked closely with OpenAI co-founder Ilya Sutskever on reinforcement learning and co-created the o1 reasoning model.

Bansal joins Meta’s ambitious superintelligence team, which is focused on further pushing AI reasoning capabilities.

Former Scale AI CEO Alexandr Wang leads the new team, brought in after Meta invested $14.3 billion in the AI data labelling company.

Alongside Bansal, several other notable figures have recently joined, including three OpenAI researchers from Zurich, a former Google DeepMind expert, Jack Rae, and a senior machine learning lead from Sesame AI.

Meta CEO Mark Zuckerberg is accelerating AI recruitment by negotiating with prominent names like former GitHub CEO Nat Friedman and Safe Superintelligence co-founder Daniel Gross.

Despite these aggressive efforts, OpenAI CEO Sam Altman revealed that even $100 million joining bonuses have failed to lure key staff away from his firm.

Zuckerberg has also explored acquiring startups such as Sutskever’s Safe SuperIntelligence and Perplexity AI, further highlighting Meta’s urgency in catching up in the generative AI race.

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

Hawaiian Airlines confirms flights are safe despite cyberattack

Hawaiian Airlines has reported a cyberattack that affected parts of its IT infrastructure, though the carrier confirmed all flights remain unaffected and are operating as scheduled.

Now part of the Alaska Air Group, the airline stated it is actively working with authorities and cybersecurity experts to investigate and resolve the incident.

In a statement, the airline stressed that the safety and security of passengers and staff remain its highest priority. It has taken steps to protect its systems, restoring affected services while continuing full operations. No disruption to passenger travel has been reported.

The exact nature of the attack has not been disclosed, and no group has claimed responsibility so far. The Federal Aviation Administration (FAA) confirmed it monitors the situation closely and remains in contact with the airline. It added that there has been no impact on flight safety.

Cyberattacks in aviation are becoming increasingly common due to the sector’s heavy reliance on complex digital systems. Earlier incidents this year included cyberattacks on WestJet and Japan Airlines, which caused operational disruptions but did not compromise passenger data.

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

Infosys chairman warns of global risks from tariffs and AI

Infosys chairman Nandan Nilekani has warned of mounting global uncertainty driven by tariff wars, AI and the ongoing energy transition.

At the company’s 44th annual general meeting, he urged businesses to de-risk sourcing and diversify supply chains as geopolitical trade tensions reshape global commerce.

He described a ‘perfect storm’ of converging challenges pushing the world away from a single global market and towards fragmented trade blocs. As firms navigate the shift, they must choose between regions and adopt more strategic, resilient supply networks.

Addressing AI, Nilekani acknowledged the disruption it may bring to the workforce but framed it as an opportunity for digital transformation. He said Infosys is investing in both ‘AI foundries’ for innovation and ‘AI factories’ for scale, with over 275,000 employees already trained in AI technologies.

Energy transition was also flagged as a significant uncertainty, as the future depends on breakthroughs in renewable sources like solar, wind and hydrogen. Nilekani stressed that all businesses now face rapid technological and operational change before they can progress confidently into an unpredictable future.

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

Google releases free Gemini CLI tool for developers

Google has introduced Gemini CLI, a free, open-source AI tool that connects developers directly to its Gemini AI models. The new agentic utility allows developers to request debugging, generate code, and run commands using natural language within their terminal environment.

Built as a lightweight interface, Gemini CLI provides a streamlined way to interact with Gemini. While its coding features stand out, Google says the tool handles content creation, deep research, and complex task management across various workflows.

Gemini CLI uses Gemini 2.5 Pro for coding and reasoning tasks by default. Still, it can also connect to other AI models, such as Imagen and Veo, for image and video generation. It supports the Model Context Protocol (MCP) and integrates with Gemini Code Assist.

Moreover, the tool is available on Windows, MacOS, and Linux, offering developers a free usage tier. Access through Vertex AI or AI Studio is available on a pay-as-you-go basis for advanced setups involving multiple agents or custom models.

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

Meta wins copyright case over AI training

Meta has won a copyright lawsuit brought by a group of authors who accused the company of using their books without permission to train its Llama generative AI.

A US federal judge in San Francisco ruled the AI training was ‘transformative’ enough to qualify as fair use under copyright law.

Judge Vince Chhabria noted, however, that future claims could be more successful. He warned that using copyrighted books to build tools capable of flooding the market with competing works may not always be protected by fair use, especially when such tools generate vast profits.

The case involved pirated copies of books, including Sarah Silverman’s memoir ‘The Bedwetter’ and Junot Diaz’s award-winning novel ‘The Brief Wondrous Life of Oscar Wao’. Meta defended its approach, stating that open-source AI drives innovation and relies on fair use as a key legal principle.

Chhabria clarified that the ruling does not confirm the legality of Meta’s actions, only that the plaintiffs made weak arguments. He suggested that more substantial evidence and legal framing might lead to a different outcome in future cases.

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

WhatsApp launches AI feature to sum up all the unread messages

WhatsApp has introduced a new feature using Meta AI to help users manage unread messages more easily. Named ‘Message Summaries’, the tool provides quick overviews of missed messages in individual and group chats, assisting users to catch up without scrolling through long threads.

The summaries are generated using Meta’s Private Processing technology, which operates inside a Trusted Execution Environment. The secure cloud-based system ensures that neither Meta nor WhatsApp — nor anyone else in the conversation — can access your messages or the AI-generated summaries.

According to WhatsApp, Message Summaries are entirely private. No one else in the chat can see the summary created for you. If someone attempts to interfere with the secure system, operations will stop immediately, or the change will be exposed using a built-in transparency check.

Meta has designed the system around three principles: secure data handling during processing and transmission, strict enforcement of protections against tampering, and provable transparency to track any breach attempt.

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

Nvidia becomes world’s most valuable company after stock surge

Nvidia shares hit an all-time high on 25 June, rising 4.3 percent to US$154.31. The stock has surged 63 percent since April, adding another US$1.5 trillion to its market value.

With a total market capitalisation of about US$3.77 trillion, Nvidia has overtaken Microsoft to become the world’s most valuable listed company.

Strong earnings and growing AI infrastructure spending by major clients — including Microsoft, Meta, Alphabet and Amazon — have reinforced investor confidence.

Nvidia’s CEO, Jensen Huang, told shareholders that demand remains strong and that the computer industry is still in the early stages of a major AI upgrade cycle.

Despite gaining 15 percent in 2025, following a 170 percent rise in 2024 and a 240 percent surge in 2023, Nvidia still appears reasonably valued. It trades at 31.5 times forward earnings, below its 10-year average and close to the Nasdaq 100 multiple, even though its projected growth rate is higher.

Analyst sentiment remains firmly bullish. Nearly 90 percent of analysts tracked by Bloomberg recommend buying the stock, which trades below their average price target.

Yet, Nvidia is less widely held among institutional investors than peers like Microsoft and Apple, indicating further room for buying as AI momentum continues into 2026.

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

AI sandboxes pave path for responsible innovation in developing countries

At the Internet Governance Forum 2025 in Lillestrøm, Norway, experts from around the world gathered to examine how AI sandboxes—safe, controlled environments for testing new technologies under regulatory oversight—can help ensure that innovation remains responsible and inclusive, especially in developing countries. Moderated by Sophie Tomlinson of the DataSphere Initiative, the session spotlighted the growing global appeal of sandboxes, initially developed for fintech, and now extending into healthcare, transportation, and data governance.

Speakers emphasised that sandboxes provide a much-needed collaborative space for regulators, companies, and civil society to test AI solutions before launching them into the real world. Mariana Rozo-Paz from the DataSphere Initiative likened them to childhood spaces for building and experimentation, underscoring their agility and potential for creative governance.

From the European AI Office, Alex Moltzau described how the EU AI Act integrates sandboxes to support safe innovation and cross-border collaboration. On the African continent, where 25 sandboxes already exist (mainly in finance), countries like Nigeria are using them to implement data protection laws and shape national AI strategies. However, funding and legal authority remain hurdles.

The workshop laid bare several shared challenges: limited resources, lack of clear legal frameworks, and insufficient participation in civil society. Natalie Cohen of the OECD pointed out that just 41% of countries trust governments to regulate new technologies effectively—a gap that sandboxes can help bridge. By enabling evidence-based experimentation and promoting transparency, they serve as trust-building tools among governments, businesses, and communities.

Despite regional differences, there was consensus that AI sandboxes—when well-designed and inclusive—can drive equitable digital innovation. With initiatives like the Global Sandboxes Forum and OECD toolkits in progress, stakeholders signalled a readiness to move from theory to practice, viewing sandboxes as more than just regulatory experiments—they are, increasingly, catalysts for international cooperation and responsible AI deployment.

Track all key moments from the Internet Governance Forum 2025 on our dedicated IGF page.

Top 7 AI agents transforming business in 2025

AI agents are no longer a futuristic concept — they’re now embedded in the everyday operations of major companies across sectors.

From customer service to data analysis, AI-powered agents transform workflows by handling tasks like scheduling, reporting, and decision-making with minimal human input.

Unlike simple chatbots, today’s AI agents understand context, follow multi-step instructions, and integrate seamlessly with business tools. Google’s Gemini Agents, IBM’s Watsonx Orchestrate, Microsoft Copilot, and OpenAI’s Operator are some tools that reshape how businesses function.

These systems interpret goals and act on behalf of employees, boosting productivity without needing constant prompts.

Other leading platforms include Amelia, known for its enterprise-grade capabilities in finance and telecom; Claude by Anthropic, focused on safe and transparent reasoning; and North by Cohere, which delivers sector-specific AI for clients like Oracle and SAP.

Many of these tools offer no-code or low-code setups, enabling faster adoption across HR, finance, customer support, and more.

While most agents aren’t entirely autonomous, they’re designed to perform meaningful work and evolve with feedback.

The rise of agentic AI marks a significant shift in workplace automation as businesses move beyond experimentation toward real-world implementation, one workflow at a time.

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