ai

EU prepares new data strategy for AI growth

The European Commission will soon launch a consultation on its upcoming Data Union Strategy, a key part of efforts to boost Europe’s leadership in AI.

The strategy, set to be published by the end of the year, aims to make it easier for businesses and public bodies to share data securely and efficiently across the EU.

The initiative supports the broader AI Continent Action Plan, expected to be unveiled this week, which seeks to encourage faster adoption of AI technologies by European companies.

Instead of relying on fragmented systems, the Commission wants to improve data access, digital infrastructure, and cloud capabilities while investing in talent and streamlining complex processes.

The plan includes the creation of AI factories where companies can train models using EU-based resources, and a separate Cloud and AI Development Act later this year will promote energy-efficient investments to support these goals.

Public feedback on the Data Union Strategy will be gathered from April to June as part of the consultation process.

Despite the ambition, the Commission acknowledges ongoing concerns such as uncertainty around international data flows and challenges accessing suitable data for generative AI.

Strict privacy laws like the GDPR, instead of enabling wider AI training, have led to frustration from major tech firms over regulatory delays in Europe.

For more information on these topics, visit diplomacy.edu.

New AI feature lets WordPress users build sites in minutes

WordPress.com has introduced a new AI website builder designed to help users create an entire website in just minutes.

Available now in early access, the feature allows anyone with a WordPress.com account to try it out free of charge. It uses a conversational interface that responds to user prompts to generate complete sites, including written content, images, colour schemes, and layouts.

Users begin by describing what kind of website they need—whether a blog, portfolio, or business site—and the AI does the rest.

The more specific the initial description, the more tailored the outcome will be. If the first version isn’t quite right, users can continue refining their site simply by chatting with the builder. Once the result is satisfactory, the website can be published directly through WordPress.com.

Currently limited to basic websites, the new tool does not yet support complex features such as ecommerce or external integrations. WordPress has indicated that more functionality is coming soon.

The generated sites remain fully customisable using the usual WordPress tools, giving users full control over editing and manual adjustments post-creation.

At launch, users get 30 free prompts before needing to choose a hosting plan, with pricing starting at $18 per month.

While similar AI tools have been introduced by platforms like Wix and Squarespace, WordPress’s version brings such technology to a significantly wider audience, given that the platform powers over 40% of all websites worldwide.

For more information on these topics, visit diplomacy.edu.

Victims of AI-driven sex crimes in Korea continue to grow

South Korea is facing a sharp rise in AI-related digital sex crimes, with deepfake pornography and online abuse increasingly affecting young women and children.

According to figures released by the Ministry of Gender Equality and Family and the Women’s Human Rights Institute, over 10,000 people sought help last year, marking a 14.7 percent increase from 2023.

Women made up more than 70 percent of those who contacted the Advocacy Center for Online Sexual Abuse Victims.

The majority were in their teens or twenties, with abuse often occurring via social media, messaging apps, and anonymous platforms. A growing portion of victims, including children under 10, were targeted due to the easy accessibility of AI tools.

The most frequently reported issue was ‘distribution anxiety,’ where victims feared the release of sensitive or manipulated videos, followed by blackmail and illegal filming.

Deepfake cases more than tripled in one year, with synthetic content often involving the use of female students’ images. In one notable incident, a university student and his peers used deepfake techniques to create explicit fake images of classmates and shared them on Telegram.

With over 300,000 pieces of illicit content removed in 2024, authorities warn that the majority of illegal websites are hosted overseas, complicating efforts to take down harmful material.

The South Korean government plans to strengthen its response by expanding educational outreach, supporting victims further, and implementing new laws to prevent secondary harm by allowing the removal of personal information alongside explicit images.

For more information on these topics, visit diplomacy.edu.

Microsoft pauses $1 billion data centre project in Ohio

Microsoft has announced it is ‘slowing or pausing’ some data centre construction projects, including a $1 billion plan in Ohio, amid shifting demand for AI infrastructure.

The company confirmed it would halt early-stage development on rural land in Licking County, near Columbus, and will repurpose two of the sites for farmland.

The decision follows Microsoft’s rapid scaling of infrastructure to meet the soaring demand for AI and cloud services, which has since softened. The company acknowledged that such large projects require continuous adaptation to align with customer needs.

While Microsoft did not specify other paused projects, it revealed the suspension of later stages of a Wisconsin data centre expansion.

The slowdown also coincides with changes in Microsoft’s partnership with OpenAI, with the two companies revising their agreement to allow OpenAI to build its own AI infrastructure. This move reflects broader trends in AI computing needs, which are expensive and energy-intensive.

Despite the pause in Ohio, Microsoft plans to invest over $80 billion in AI infrastructure this fiscal year, continuing its global expansion, though it will now strategically pace its growth to align with evolving business priorities.

Local officials in Licking County expressed their disappointment, as the area had been a hub for significant tech investments, including those from Google and Meta.

For more information on these topics, visit diplomacy.edu.

Google unveils new AI agent toolkit

This week at Google Cloud Next in Las Vegas, Google revealed its latest push into ‘agentic AI’. A software designed to act independently, perform tasks, and communicate with other digital systems.

Central to this effort is the Agent Development Kit (ADK), an open-source toolkit said to let developers build AI agents in under 100 lines of code.

Instead of requiring complex systems, the ADK includes pre-built connectors and a so-called ‘agent garden’ to streamline integration with data platforms like BigQuery and AlloyDB.

Google also introduced a new Agent2Agent (A2A) protocol, aimed at enabling cooperation between agents from different vendors. With over 50 partners, including Accenture, SAP and Salesforce, already involved, the company hopes to establish a shared standard for AI interaction.

Powering these tools is Google’s latest AI chip, Ironwood, a seventh-generation TPU promising tenfold performance gains over earlier models. These chips, designed for use with advanced models like Gemini 2.5, reflect Google’s ambition to dominate AI infrastructure.

Despite the buzz, analysts caution that the hype around AI agents may outpace their actual utility. While vendors like Microsoft, Salesforce and Workday push agentic AI to boost revenue, in some cases even replacing staff, experts argue that current models still fall short of real human-like intelligence.

Instead of widespread adoption, businesses are expected to focus more on managing costs and complexity, especially as economic uncertainty grows. Without strong oversight, these tools risk becoming costly, unpredictable, and difficult to scale.

For more information on these topics, visit diplomacy.edu.

Google pushes AI limits with Ironwood

Google has announced Ironwood, its latest and most advanced AI processor, marking the seventh generation of its custom Tensor Processing Unit (TPU) architecture.

Designed specifically for the growing demands of its Gemini models, particularly those requiring complex simulated reasoning, which Google refers to as ‘thinking’, Ironwood represents a significant leap forward in performance.

Instead of relying solely on software updates, Google is highlighting how hardware like Ironwood plays a central role in boosting AI capabilities, ushering in what it calls the ‘age of inference.’

However, this TPU is not just faster but dramatically more scalable. Ironwood chips will operate in tightly connected clusters of up to 9,216 units, each cooled by liquid and linked through an enhanced Inter-Chip Interconnect.

These chips can also be deployed in smaller 256-chip servers, offering flexibility for cloud developers and researchers.

Instead of offering modest improvements, Ironwood delivers a peak throughput of 4,614 teraflops per chip, alongside 192GB of memory and 7.2 terabits per second of bandwidth, making it vastly superior to its predecessor, Trillium.

Google says this advancement is more than a performance boost, it’s a foundation for building AI agents that can act on a user’s behalf by gathering information and producing outputs proactively.

Rather than functioning as passive tools, AI systems powered by Ironwood are intended to behave more independently, reflecting a growing trend toward what Google calls ‘agentic AI.’

While Google’s comparison to supercomputers like El Capitan may be flawed due to differing hardware standards, there’s no doubt Ironwood is a substantial upgrade. The company claims it is twice as powerful per watt as the v5p TPU, even if the newer Trillium (v6) chip wasn’t included in the comparison.

Regardless, Ironwood is expected to power the next generation of AI breakthroughs, as the company prepares to move beyond its current Gemini 2.5 model.

For more information on these topics, visit diplomacy.edu.

Gemini 2.5 Pro boosts Deep Research tool with smarter AI

Google has upgraded its Deep Research tool with the experimental Gemini 2.5 Pro model, promising major improvements in how users access and process complex information.

Deep Research acts as an AI research assistant capable of scanning hundreds of websites, evaluating content, and producing multi-page reports complete with citations and even podcast-style summaries.

Previously powered by Gemini 2.0 Flash, the new iteration significantly enhances reasoning, planning, and reporting capabilities. Human evaluators in Google’s testing preferred Deep Research’s outputs over those generated by OpenAI’s equivalent by a ratio greater than 2 to 1.

Users also noted clearer analytical thinking and better synthesis of information across sources.

The Gemini 2.5 Pro upgrade is available now to Gemini Advanced subscribers across web, Android, and iOS platforms.

For those using the free version, the Gemini 2.0 Flash model remains accessible in over 150 countries, continuing Google’s push to offer powerful research tools to a wide user base.

For more information on these topics, visit diplomacy.edu.

Blockchain app ARK fights to keep human creativity ahead of AI

Nearly 20 years after his AI career scare, screenwriter Ed Bennett-Coles and songwriter Jamie Hartman have developed ARK, a blockchain app designed to safeguard creative work from AI exploitation.

The platform lets artists register ownership of their ideas at every stage, from initial concept to final product, using biometric security and blockchain verification instead of traditional copyright systems.

ARK aims to protect human creativity in an AI-dominated world. ‘It’s about ring-fencing the creative process so artists can still earn a living,’ Hartman told AFP.

The app, backed by Claritas Capital and BMI, uses decentralised blockchain technology instead of centralised systems to give creators full control over their intellectual property.

Launching summer 2025, ARK challenges AI’s ‘growth at all costs’ mentality by emphasising creative journeys over end products.

Bennett-Coles compares AI content to online meat delivery, efficient but soulless, while human artistry resembles a grandfather’s butcher trip, where the experience matters as much as the result.

The duo hopes their solution will inspire industries to modernise copyright protections before AI erodes them completely.

For more information on these topics, visit diplomacy.edu.

Microsoft’s Copilot Vision now sees your entire screen to guide you through apps

Microsoft is testing a major upgrade to its Copilot AI that can view your entire screen instead of just working within the Edge browser.

The new Copilot Vision feature helps users navigate apps like Photoshop and Minecraft by analysing what’s on display and offering step-by-step guidance, even highlighting specific tools instead of just giving verbal instructions.

The feature operates more like a shared Teams screen instead of Microsoft’s controversial Recall snapshot system.

Currently limited to US beta testers, Copilot Vision will eventually highlight interface elements directly on users’ screens. It works on standard Windows PCs instead of requiring specialised Copilot+ hardware, with mobile versions coming to iOS and Android.

Alongside visual assistance, Microsoft is adding document search capabilities. Copilot can now find information within files like Word documents and PDFs instead of just searching by filename.

Both updates will roll out fully in the coming weeks, potentially transforming how users interact with both apps and documents on their Windows devices.

For more information on these topics, visit diplomacy.edu.

Amazon launches Nova Sonic AI for natural voice interactions

Amazon has unveiled Nova Sonic, a new AI model designed to process and generate human-like speech, positioning it as a rival to OpenAI and Google’s top voice assistants. The company claims it outperforms competitors in speed, accuracy, and cost, and it is reportedly 80% cheaper than GPT-4o.

Already powering Alexa+, Nova Sonic excels in real-time conversation, handling interruptions and noisy environments better than legacy AI assistants.

Unlike older voice models, Nova Sonic can dynamically route requests, fetching live data or triggering external actions when needed. Amazon says it achieves a 4.2% word error rate across multiple languages and responds in just 1.09 seconds, faster than OpenAI’s GPT-4o.

Developers can access it via Bedrock, Amazon’s AI platform, using a new streaming API.

The launch signals Amazon’s push into artificial general intelligence (AGI), AI that mimics human capabilities.

Rohit Prasad, head of Amazon’s AGI division, hinted at future models handling images, video, and sensory data. This follows last week’s preview of Nova Act, an AI for browser tasks, suggesting Amazon is accelerating its AI rollout beyond Alexa.

For more information on these topics, visit diplomacy.edu.