Artificial intelligence

AI teachers and deepfakes tested to ease UK teacher shortages

Amid a worsening recruitment and retention crisis in UK education, some schools are trialling AI-based teaching solutions, including remote teachers delivered via video links and even proposals for deepfake avatars to give lessons.

These pilots are part of efforts to maintain educational provision where qualified staff are scarce, with proponents arguing that technology can help reduce teacher workload and address gaps in core subjects, such as mathematics.

However, many teachers and unions remain sceptical or critical. Some educators argue that remote or AI-led instruction cannot replace the human presence, interpersonal support and contextual knowledge provided by in-room teachers.

Union activity and petitions opposing virtual teaching arrangements reflect broader concerns about the implications for job security, education quality and the potential de-professionalisation of teaching.

The BBC’s reporting highlighted specific examples, such as a Lancashire secondary school bringing in a remote maths teacher based hundreds of miles away, a move that sparked debate among local teachers who emphasise the irreplaceable role of in-person interaction in effective teaching.

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

ChatGPT tops Apple’s 2025 app downloads in the US

Apple has released its annual ranking of the most downloaded apps and games, with ChatGPT taking the top spot among free iPhone apps in the United States for 2025, marking a major moment for AI in mainstream consumer use.

The OpenAI chatbot rose from fourth place last year, surpassing established social platforms and everyday utilities. Its ascent highlights how quickly AI-driven tools have become embedded in daily habits and how they may challenge the dominance of traditional search apps on mobile devices.

Apple’s charts show broader shifts across categories. Threads, Google, TikTok, WhatsApp, and Instagram also ranked highly among free iPhone downloads, while Google’s Gemini entered the top ten, reflecting the growing presence of competing AI assistants in the mobile ecosystem.

Gaming trends remained strong. Block Blast! led the US free iPhone games list, while Minecraft held its position as the top paid title across devices. ChatGPT also became the second-most downloaded free app on iPad, signalling consistent demand for AI across screens.

Apple says the rankings reflect the evolving mix of entertainment, creativity, and productivity tools shaping the App Store landscape, as AI continues to influence how people search, work, and play across its platforms.

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

DeVry improves student support with AI

In the US, DeVry University has upgraded its student support system by deploying Salesforce Agentforce 360, aiming to offer faster and more personalised assistance to its 32,000 learners.

The new AI agents provide round-the-clock support for DeVryPro, the university’s online learning programme, ensuring students receive timely guidance.

The platform also simplifies course enrolment through a self-service website, allowing learners to manage enrolment and payments efficiently. Real-time guidance replaces the previous chatbot, helping students access course information and support outside regular hours.

With Data 360 integrating information from multiple systems, DeVry can deliver personalised recommendations while automating time-consuming tasks such as weekly onboarding.

Advisors can now focus on building stronger connections with students and supporting the development of workforce skills.

University leaders emphasise that these advancements reflect a commitment to preparing learners for an AI-driven workforce, combining innovative technology with personalised academic experiences. The initiative positions DeVry as a leader in integrating AI into higher education.

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

Multimodal AI reveals new immune patterns across cancer types

A recent study examined the capabilities of GigaTIME, a multimodal AI framework that models the tumour immune microenvironment by converting routine H and E slides into virtual multiplex immunofluorescence images.

Researchers aimed to solve long-standing challenges in profiling tumour ecosystems by using a scalable and inexpensive technique instead of laboratory methods that require multiple samples and extensive resources.

The study focused on how large image datasets could reveal patterns of protein activity that shape cancer progression and therapeutic response.

GigaTIME was trained on millions of matched cells and applied to more than fourteen thousand slides drawn from a wide clinical network. The system generated nearly 300.000 virtual images and uncovered over 1000 associations between protein channels and clinical biomarkers.

Spatial features such as sharpness, entropy and signal variability were often more informative than density alone, revealing immune interactions that differ strongly across cancer types.

When tested on external tumour collections, the framework maintained strong performance and consistently exceeded the results of comparator models.

The study reported that GigaTIME could identify patterns linked to tumour invasion, survival and stage. Protein combinations offered a clearer view of immune behaviour than single markers, and the virtual signatures aligned with known and emerging genomic alterations.

Certain proteins were easier to infer than others, which reflected structural differences at the cellular level rather than model limitations. The research also suggested that immune evasion mechanisms may shift during advanced disease, altering how proteins such as PD-L1 contribute to tumour progression.

The authors argued that virtual multiplex imaging could expand access to spatial proteomics for both research and clinical practice.

Wider demographic representation and broader protein coverage are necessary for future development, yet the approach demonstrated clear potential to support large population studies instead of the restricted datasets produced through traditional staining methods.

Continued work seeks to build a comprehensive atlas and refine cell-level segmentation to deepen understanding of immune and tumour interactions.

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

Pinterest’s generative tools showcased in AWS report

AWS has released a new case study showing how Pinterest scaled into an AI-powered discovery platform, emphasising the cloud provider’s role in supporting rapid growth and responsible use of generative tools across the service.

AWS says Pinterest now serves around 600 million users and relies on cloud infrastructure to process large visual datasets. Its systems analyse terabytes of content each day and generate high volumes of personalised suggestions across search, shopping, and inspiration features.

The case study details Pinterest’s move from early machine-learning models to multimodal and generative systems built on AWS. It highlights Canvas for image enhancement, improved visual search, and the conversational Pinterest Assistant.

AWS also points to Pinterest’s use of Amazon EKS, EC2 GPU instances, and Bedrock-powered moderation tools as a full-stack approach to responsible AI. Pinterest states that these systems help maintain a safe and positive environment while supporting new commercial and creative features.

AWS cites recent performance metrics as evidence of effective scaling, noting gains in revenue, user activity, and search quality. The company presents the case study as evidence that cloud-based AI infrastructure can support innovation on a global scale.

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

Big Tech boosts India’s AI ambitions amid concerns over talent flight and limited infrastructure

Major announcements from Microsoft ($17.5bn) and Amazon (over $35bn by 2030) have placed India at the centre of global AI investment trends, offering momentum at a time when analysts frame Indian markets as a ‘hedge’ against a potential global AI bubble.

While India has rapidly adopted AI and attracted substantial funding for data centres and chip manufacturing, including a new collaboration between Intel and Tata Electronics, the country remains a follower rather than a frontrunner in sovereign AI capabilities.

India’s government is preparing to launch its first sovereign AI model, which will support more than 22 languages. Yet its $1.25 billion investment is dwarfed by France’s €117 billion and Saudi Arabia’s $100 billion AI programmes, leaving India far behind in compute availability, R&D depth, and semiconductor infrastructure.

Despite having 2.5 times the global average concentration of AI-skilled professionals, the country faces persistent talent flight due to limited high-end domestic opportunities and a lack of competitive policy incentives.

According to EY and UNCTAD, India’ punches above its weight’ relative to its economic stage, ranking among the top nations in AI talent, startup activity, and scientific publications. Still, funding gaps remain stark: Indian AI startups raised just $1.16 billion, compared to more than $100 billion in the US and nearly $10 billion in China.

India’s emerging strength lies less in foundation-model development and more in downstream AI applications, where cost-efficient tools can drive entrepreneurship and solve local challenges such as agriculture, education, and public service delivery. Apps like MahaVISTAAR, reaching over 15 million farmers, illustrate this direction.

Yet AI also poses a threat to India’s economic backbone. Analysts warn that the country’s IT services sector, which has long been a pillar of growth, is becoming increasingly vulnerable as AI automates core business functions. Underperformance in IT stocks, reduced hiring, and stagnant wages signal early disruption.

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

Adobe brings its leading creative tools straight into ChatGPT

Yesterday, Adobe opened a new chapter for digital creativity by introducing Photoshop, Adobe Express and Adobe Acrobat inside ChatGPT.

The integration gives 800 million weekly users direct access to trusted creative and productivity tools through a conversational interface. Adobe aims to make creative work easier for newcomers by linking its technology to simple written instructions.

Photoshop inside ChatGPT offers selective edits, tone adjustments and creative effects, while Adobe Express brings quick design templates and animation features to people who want polished content without switching between applications.

Acrobat adds powerful document controls, allowing users to organise, edit or redact PDFs inside the chat. Each action blends conversation with Adobe’s familiar toolsets, giving users either simple text-driven commands or fine control through intuitive sliders.

The launch reflects Adobe’s broader investment in agentic AI and its Model Context Protocol. Earlier releases such as Acrobat Studio and AI Assistants for Photoshop and Adobe Express signalled Adobe’s ambition to expand conversational creative experiences.

Adobe also plans to extend an upcoming Firefly AI Assistant across multiple apps to support faster movement from an idea to a finished design.

All three apps are now available to ChatGPT users on desktop, web and iOS, with Android support expanding soon. Adobe positions the integration as an entry point for new audiences who may later move into the full desktop versions for deeper control.

The company expects the partnership to widen access to creative expression by letting anyone edit images, produce designs or transform documents simply by describing what they want to achieve.

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

Three in ten US teens now use AI chatbots every day, survey finds

According to new data from the Pew Research Center, roughly 64% of US teens (aged 13–17) say they have used an AI chatbot; about three in ten (≈ 30%) report daily use. Among those teens, the leading chatbot is ChatGPT (used by 59%), followed by Gemini (23%) and Meta AI (20%).

The widespread adoption raises growing safety and welfare concerns. As teenagers increasingly rely on AI for information, companionship or emotional support, critics point to potential risks, including exposure to biased content, misinformation, or emotionally manipulative interactions, particularly among vulnerable youth.

Legal action has already followed, with families of at least two minors suing AI-developer companies after alleged harmful advice from chatbots.

Demographic patterns reveal that Black and Hispanic teens report higher daily usage rates (around 33-35%) compared to their White peers (≈ 22%). Daily use is also more common among older teens (15–17) than younger ones.

For policymakers and digital governance stakeholders, the findings add urgency to calls for AI-specific safeguarding frameworks, especially where young people are concerned. As AI tools become embedded in adolescent life, ensuring transparency, responsible design, and robust oversight will be critical to preventing unintended harms.

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

AI model predicts prediabetes with high accuracy

Researchers have developed an AI model that enhances prediabetes prediction by integrating oxidative stress markers with traditional clinical indicators. The Pattern Neural Network model achieved 98.3% accuracy in Indian adults, outperforming other machine learning methods.

Total antioxidant status emerged as a key predictor, with lower antioxidant capacity observed in individuals with prediabetes. Waist circumference and BMI were also highly informative, alongside glucose markers such as HbA1c and OGTT.

The inclusion of oxidative stress measures provides a deeper understanding of the mechanisms underlying metabolic risk.

The study used clinical and biochemical data from 199 adults, with the PNN trained on 14 features, including demographic and biochemical variables. High accuracy across all sets indicates strong potential for quick, low-cost screening and personalised early interventions.

While the results are promising, the single-centre design and limited sample size indicate that external validation is needed. Future studies should test the model in larger, multi-site cohorts and integrate longitudinal data to enhance its real-world applicability and public health impact.

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

EU proposes easing environmental rules for datacentres and housing

The European Commission has proposed a significant overhaul of environmental rules, potentially exempting datacentres, AI facilities, and affordable housing from mandatory impact assessments.

Member states would retain discretion over whether such projects require full environmental scrutiny, as part of a broader plan to expedite permitting and reduce reporting obligations for businesses.

The package also repeals a hazardous chemical database, eases polluter obligations, and moves environmental management rules from individual plants to whole companies. The commission states that the changes could save firms €1 billion annually, but green groups warn of potential costs to health and biodiversity.

The proposals align with plans to modernise the EU electricity grid and new climate targets to reduce emissions by 90% compared to 1990. Experts have cautioned that loopholes allowing international carbon credits could weaken domestic emissions reductions.

Corporate sustainability laws are also being scaled back. The revised rules limit the number of companies covered, postpone compliance deadlines to 2029, and remove obligations to implement climate transition plans.

Business lobby groups have welcomed the changes as a more realistic approach to corporate social responsibility and due diligence.

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