Artificial intelligence

Large language models mimic human object perception

Recent research shows that large multimodal language models (LLMs) can develop object representations strikingly similar to human cognition. By analysing how these AI models understand and organise concepts, scientists found patterns in the models that mirror neural activity in the human brain.

The study examined embeddings for 1,854 natural objects, derived from millions of text-image pairings. These embeddings capture relationships between objects and were compared with brain scan data from regions like EBA, PPA, RSC and FFA.

Researchers also discovered that multimodal training, which combines text and image data, enhances model’s ability to form these human-like concepts. Findings suggest that large language models can achieve more natural understanding of the world, offering potential improvements in human-AI interaction and future model design.

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

UC Santa Cruz uses NVIDIA AI to map global coastal flood risks

Researchers at the University of California, Santa Cruz, are using NVIDIA’s accelerated computing to model coastal flooding and support climate adaptation planning.

Led by Professor Michael Beck, the team develops high-resolution, GPU-powered visualisations to assess how coral reefs, mangroves, and dunes can reduce flood damage.

The centre employs NVIDIA CUDA-X software and RTX GPUs to speed up flood simulations from six hours to just 40 minutes. Using tools such as SFINCS and Unreal Engine 5, the team can now generate interactive visual models of storm impact scenarios, providing vital insights for governments and insurers.

The researchers’ current goal is to map flooding risks across small island states worldwide ahead of COP30. Their previous visualisations have already helped secure reef insurance policies in Mexico’s Mesoamerican Barrier Reef region, ensuring funding for coral restoration after severe storms.

A project, part of CoSMoS ADAPT, that aims to expand the US Geological Survey’s coastal modelling system and integrate nature-based solutions like dunes and reefs into large-scale flood resilience strategies.

Through NVIDIA’s technology and academic grants, the initiative demonstrates how accelerated computing can drive real-world environmental protection.

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

Federal Reserve embraces crypto innovation in payments system

The Federal Reserve has signalled a shift towards decentralised finance, with Governor Waller saying the central bank now welcomes crypto innovators into mainstream payments.

Speaking at the Payments Innovation Conference on 21 October, Waller said the Fed intends to play an active role in the ongoing technology-driven transformation of the financial system.

Waller highlighted how stablecoins, tokenised assets, and AI are reshaping the payments landscape. He said private firms drive innovation but added that public institutions like the Fed must adapt to support evolving financial systems.

The governor said the central bank is exploring how tokenisation, smart contracts, and AI could enhance its own systems and foster closer dialogue with industry innovators.

In a significant policy proposal, Waller revealed that the Fed is studying a new type of ‘payment account’ for legally eligible institutions. The concept would provide streamlined access to Federal Reserve payment rails for fintech and crypto firms without requiring a full master account.

Such accounts would operate under tighter controls, including balance caps, no interest payments, and no overdraft privileges, allowing faster review times while maintaining system safety.

Waller said the payments revolution is underway and urged collaboration between traditional finance and emerging digital sectors. He called the event a turning point for Fed–innovator relations, noting that crypto and distributed ledgers are now part of modern payments.

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

EU pushes harder on basic digital skills for growth

Nearly half of EU adults lack basic digital skills, yet most jobs demand them. Eurostat reports only 56% have at least basic proficiency. EU Code Week spotlights the urgency for digital literacy and inclusion.

The Digital Education Action Plan aims to modernise curricula, improve infrastructure, and train teachers. EU policymakers target 80% of adults with basic skills by 2030. Midway progress suggests stronger national action is still required.

Progress remains uneven across regions, with rural connectivity still lagging in places. Belgium began a school smartphone ban across Flanders from 1 September to curb distractions. Educators now balance classroom technology with attention and safety.

Brussels proposed a Union of Skills strategy to align education and competitiveness. The EU also earmarked fresh funding for AI, cybersecurity, and digital skills. Families and schools are urged to develop unplugged problem-solving alongside classroom learning.

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

Deepfake targeting Irish presidential candidate sparks election integrity warning

Irish presidential candidate Catherine Connolly condemned a deepfake AI video that falsely announced her withdrawal from the race. The clip, designed to resemble an RTÉ News broadcast, spread online before being reported and removed from major social media platforms.

Connolly said the video was a disgraceful effort to mislead voters and damage democracy. Her campaign team filed a complaint with the Irish Electoral Commission and requested that all copies be clearly labelled as fake.

Experts at Dublin City University identified slight distortions in speech and lighting as signs of AI manipulation. They warned that the rapid spread of synthetic videos underscores weak content moderation by online platforms.

Connolly urged the public not to share the clip and to respond through civic participation. Authorities are monitoring digital interference as Ireland prepares for its presidential vote on Friday.

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

General Motors unveils AI and robotics innovations for next-generation vehicles

General Motors showcased AI, robotics, and software at GM Forward, showing how cars are becoming intelligent assistants. CEO Mary Barra unveiled advances in autonomy, AI, computing, and energy, signalling a move toward smarter, safer, and more personalised vehicles.

GM plans to bring eyes-off driving to market in 2028, debuting on the Cadillac ESCALADE IQ electric SUV. The company has already mapped 600,000 miles of hands-free roads in North America, with Super Cruise drivers completing 700 million miles without a crash attributed to the system.

Cruise’s driverless technology adds over five million fully autonomous miles of experience, forming the foundation for GM’s next phase of personal autonomy.

Vehicles will also feature conversational AI powered by Google Gemini from next year, allowing drivers to interact naturally with their car. In the future, GM plans to introduce its own AI, fine-tuned to individual vehicles and driver preferences via OnStar.

A new centralised computing platform, launching in 2028, will integrate propulsion, steering, infotainment, and safety on a single high-speed core, increasing update capacity, bandwidth, and AI performance dramatically.

GM is also advancing robotics at its Autonomous Robotics Center and Mountain View lab. Collaborative robots trained on decades of production data are being deployed in US assembly plants to improve efficiency, safety, and workplace quality.

Additionally, GM’s EVs can provide backup power to homes and, from 2026, to the electrical grid, supported by an integrated home energy system and smartphone app. Together, these innovations mark a significant step toward intelligent, personalised, and sustainable mobility.

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

Amazon launches Blue Jay and Project Eluna to support employees

Amazon has unveiled two new innovations, Blue Jay and Project Eluna, designed to improve efficiency and safety in its operations. Blue Jay coordinates multiple arms to handle items in one workspace, reducing repetitive tasks and supporting employees.

Project Eluna is an agentic AI model that helps operators make data-driven decisions, anticipating bottlenecks and optimising workflows.

Blue Jay uses robotics experience, AI, and digital twin simulations to go from concept to production in just over a year. It is being tested in South Carolina, managing 75% of items and could support Amazon’s Same-Day delivery network.

Project Eluna will pilot in Tennessee, offering operators clear recommendations and reducing the cognitive load of monitoring multiple dashboards.

These systems aim to enhance the employee experience by improving ergonomics, reducing repetitive tasks, and opening new career pathways. Amazon is expanding robotics, mechatronics, and AI training so employees can work confidently with these technologies.

Blue Jay and Project Eluna join other recent innovations, including Vulcan, a robot with a sense of touch, and DeepFleet, an AI model coordinating fleets of mobile robots.

Tye Brady, Amazon Robotics chief technologist, emphasised that the focus remains on people. AI and robotics integration aims to enhance workplace safety, efficiency, and fulfillment, reflecting Amazon’s focus on workforce development and technological progress.

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

UK data stays in the UK as OpenAI rolls out residency

OpenAI will offer UK data residency for API Platform, ChatGPT Enterprise, and ChatGPT Edu from October 24. The option, announced by Deputy PM David Lammy, is tied to a Ministry of Justice partnership. The government says it boosts privacy, security, and resilience for public services and business.

Lammy will unveil the ‘sovereign capability’ at OpenAI Frontiers, citing early MoJ efficiency gains. Over 1,000 probation officers will use Justice Transcribe to record and auto-transcribe offender meetings. Hours of admin shift to AI so staff can focus on supervision and public protection.

OpenAI CEO Sam Altman says UK usage has quadrupled in the past year. The company pitches AI as a way to save time and lift productivity across sectors. MoJ pilots have sparked interest from other departments, with broader adoption expected.

Data residency is a key blocker for regulated sectors, and this move aims to address that gap. Keeping data within the UK can simplify compliance and reduce perceived risk. It also underpins continuity plans by localising sensitive workloads.

ChatGPT Atlas, an AI-first web browser, was also announced this week. Its arrival could nudge users away from keyword searches toward conversational answers. OpenAI faces rivals Anthropic, Perplexity, and big tech incumbents in that shift.

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

AI leaders call for a global pause in superintelligence development

More than 850 public figures, including leading computer scientists Geoffrey Hinton and Yoshua Bengio, have signed a joint statement urging a global slowdown in the development of artificial superintelligence.

The open letter warns that unchecked progress could lead to human economic displacement, loss of freedom, and even extinction.

An appeal that follows growing anxiety that the rush toward machines surpassing human cognition could spiral beyond human control. Alan Turing predicted as early as the 1950s that machines might eventually dominate by default, a view that continues to resonate among AI researchers today.

Despite such fears, global powers still view the AI race as essential for national security and technological advancement.

Tech firms like Meta are also exploiting the superintelligence label to promote their most ambitious models, while leaders such as OpenAI’s Sam Altman and Microsoft’s Mustafa Suleyman have previously acknowledged the existential risks of developing systems beyond human understanding.

The statement calls for an international prohibition on superintelligence research until there is a broad scientific consensus on safety and public approval.

Its signatories include technologists, academics, religious figures, and cultural personalities, reflecting a rare cross-sector demand for restraint in an era defined by rapid automation.

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

ChatGPT faces EU’s toughest platform rules after 120 million users

OpenAI’s ChatGPT could soon face the EU’s strictest platform regulations under the Digital Services Act (DSA), after surpassing 120 million monthly users in Europe.

A milestone that places OpenAI’s chatbot above the 45 million-user threshold that triggers heightened oversight.

The DSA imposes stricter obligations on major platforms such as Meta, TikTok, and Amazon, requiring greater transparency, risk assessments, and annual fees to fund EU supervision.

The European Commission confirmed it has begun assessing ChatGPT’s eligibility for the ‘very large online platform’ status, which would bring the total number of regulated platforms to 26.

OpenAI reported that its ChatGPT search function alone had 120.4 million monthly active users across the EU in the six months ending 30 September 2025. Globally, the chatbot now counts around 700 million weekly users.

If designated under the DSA, ChatGPT would be required to curb illegal and harmful content more rigorously and demonstrate how its algorithms handle information, marking the EU’s most direct regulatory test yet for generative AI.

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