ai

Microsoft brings Grok AI to Azure

Microsoft has become one of the first major cloud providers to offer managed access to Grok, the controversial AI model from Elon Musk’s xAI startup.

Now available through the Azure AI Foundry platform, both Grok 3 and Grok 3 mini will be billed by Microsoft and include the same service-level agreements as other Azure-hosted models.

Grok gained attention for its unfiltered and provocative tone, marketed by Musk as a more candid alternative to mainstream AI.

Unlike ChatGPT, it has been known to use vulgar language and provide responses on sensitive topics that other models typically avoid.

However, the AI has stirred criticism, particularly over troubling behaviour such as undressing women in photos and referencing conspiracy theories. Incidents of censorship and offensive content have raised concerns about its deployment on Musk’s platform X.

Instead of replicating that experience, Microsoft is offering a more controlled version of Grok within Azure. These versions include stricter content controls, enhanced data integration, and improved governance tools, distinguishing them from the models directly available through xAI.

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

Can AI replace therapists?

With mental health waitlists at record highs and many struggling to access affordable therapy, some are turning to AI chatbots for support.

Kelly, who waited months for NHS therapy, found solace in character.ai bots, describing them as always available, judgment-free companions. ‘It was like a cheerleader,’ she says, noting how bots helped her cope with anxiety and heartbreak.

But despite emotional benefits for some, AI chatbots are not without serious risks. Character.ai is facing a lawsuit from the mother of a 14-year-old who died by suicide after reportedly forming a harmful relationship with an AI character.

Other bots, like one from the National Eating Disorder Association, were shut down after giving dangerous advice.

Even so, demand is high. In April 2024 alone, 426,000 mental health referrals were made in England, and over a million people are still waiting for care. Apps like Wysa, used by 30 NHS services, aim to fill the gap by offering CBT-based self-help tools and crisis support.

Experts warn, however, that chatbots lack context, emotional intuition, and safeguarding. Professor Hamed Haddadi calls them ‘inexperienced therapists’ that may agree too easily or misunderstand users.

Ethicists like Dr Paula Boddington point to bias and cultural gaps in the AI training data. And privacy is a looming concern: ‘You’re not entirely sure how your data is being used,’ says psychologist Ian MacRae.

Still, users like Nicholas, who lives with autism and depression, say AI has helped when no one else was available. ‘It was so empathetic,’ he recalls, describing how Wysa comforted him during a night of crisis.

A Dartmouth study found AI users saw a 51% drop in depressive symptoms, but even its authors stress bots can’t replace human therapists. Most experts agree AI tools may serve as temporary relief or early intervention—but not as long-term substitutes.

As John, another user, puts it: ‘It’s a stopgap. When nothing else is there, you clutch at straws.’

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

Baby Boomers and Gen X embrace AI shopping

New research from Adyen shows that over a third of global consumers now use AI to assist with shopping. In a survey of 41,000 people across 28 countries, 12 per cent reported using AI for the first time in the past year, and more than half expressed willingness to continue.

Consumers praised AI for providing inspiration and helping them discover unique brands. The fastest growth in AI use came from Baby Boomers and Generation X, though younger groups still dominate in actual usage.

While 57 per cent of Gen Z shoppers rely on AI, just 16 per cent of Boomers currently do so, despite a sharp rise in adoption. Retailers, in turn, are investing in AI to boost sales, product innovation, and fraud prevention.

Adyen’s report also highlights the need for better integration between online and offline shopping experiences, with many retailers still lacking unified commerce offerings. Despite digital growth, physical stores remain the preferred channel for most shoppers, who value immediacy and tactile experiences.

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

AI microscope detects live blood clot activity

Researchers at the University of Tokyo have developed an AI-powered microscope that can observe blood clot formation in real time.

By capturing thousands of images per second, the tool analyses how platelets behave and clump together, offering a non-invasive way to assess clot risk in patients with coronary artery disease.

Unlike current methods that rely on indirect markers or invasive procedures, the AI tool provides live visual data using a simple blood draw from the arm. The results closely matched samples taken directly from the arteries, making the approach both safer and more accessible.

Japanese doctors could soon use this technology to evaluate how well blood-thinning drugs are working on an individual basis. While still in the research phase, experts believe it may eventually transform point-of-care decision-making for heart patients.

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

OpenAI launches advanced coding assistant Codex

OpenAI has launched Codex, a new AI coding agent designed to streamline software development by automating routine tasks and improving code reliability.

Built on a version of its o3 model known as codex-1, the agent uses reinforcement learning to generate high-quality code and test it before output.

Codex operates in a secure, cloud-based sandbox that mirrors a user’s environment and integrates with GitHub for real-time access to repositories.

It logs every step, provides test results, and supports customisation through AGENTS.md files, allowing developers to guide the AI.

Currently available to ChatGPT Pro, Enterprise, and Team subscribers, Codex is being piloted by major firms like Cisco, Superhuman, and Kodiak.

OpenAI plans wider access and future upgrades for more complex, asynchronous collaboration, though limitations like lack of image input support remain.

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

AI Darth Vader in Fortnite sparks union dispute

The use of an AI-generated Darth Vader voice in Fortnite has triggered a legal dispute between SAG-AFTRA and Epic Games.

According to GamesIndustry.biz, the actors’ union filed an unfair labor practice complaint, claiming it was not informed or consulted about the decision to use an artificial voice model in the game.

In Fortnite’s Galactic Battle season, players who defeat Darth Vader in Battle Royale can recruit him, triggering limited voice interactions powered by conversational AI.

The voice used stems from a licensing agreement with the estate of James Earl Jones, who retired in 2022 and granted rights for AI use of his iconic performance.

While Epic Games has confirmed it had legal permission to use Jones’ voice, SAG-AFTRA alleges the company bypassed union protocols by not informing them or offering the role to a human actor.

The outcome of this dispute could have broader implications for how AI voices are integrated into video games and media going forward, particularly regarding labor rights and union oversight.

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

Google releases NotebookLM app early

Google has launched its AI-powered research assistant, NotebookLM, on Android and iOS a day earlier than expected and just ahead of its annual I/O developer conference.

Until now, the service was only available on desktop, but users can now access its full features while on the move.

NotebookLM helps users understand complex content by offering intelligent summaries and allowing them to ask questions directly about their documents.

A standout feature, Audio Overviews, creates AI-generated podcast-style summaries from uploaded materials and supports offline listening and background playback.

Mobile users can now create and manage notebooks directly from their devices. Instead of limiting content sources, the app enables users to add websites, PDFs, or YouTube videos by simply tapping the share icon and selecting NotebookLM.

It also offers easy access to previously added sources and adapts its appearance to match the device’s light or dark mode settings.

With the release timed just before Google’s keynote, it’s likely the company will highlight NotebookLM’s capabilities further during the I/O 2025 presentation.

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

Taiwan boosts AI power with Nvidia supercomputer

Nvidia and Foxconn have announced plans to build a major AI factory supercomputer in Taiwan, backed by the Taiwanese government.

The facility, powered by 10,000 of Nvidia’s Blackwell GPUs, will be delivered through Foxconn’s Big Innovation Company and will support researchers, start-ups, and industry.

Taiwan’s National Science and Technology Council will use the platform to provide real-time AI cloud computing resources across its tech ecosystem. TSMC researchers also plan to utilise the infrastructure to drive innovation in semiconductor development.

The announcement highlights Nvidia’s growing ties with Taiwanese companies and reflects the strategic importance of AI to Taiwan’s technological ambitions. Leaders from Nvidia, Foxconn, and TSMC described the project as a step towards creating a smarter, AI-enabled island economy.

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

Lords reject UK AI copyright bill again

The UK government has suffered a second defeat in the House of Lords over its Data (Use and Access) Bill, as peers once again backed a copyright-focused amendment aimed at protecting artists from AI content scraping.

Baroness Kidron, a filmmaker and digital rights advocate, led the charge, accusing ministers of listening to the ‘sweet whisperings of Silicon Valley’ and allowing tech firms to ‘redefine theft’ by exploiting copyrighted material without permission.

Her amendment would force AI companies to disclose their training data sources and obtain consent from rights holders.

The government had previously rejected this amendment, arguing it would lead to ‘piecemeal’ legislation and pre-empt ongoing consultations.

But Kidron’s position was strongly supported across party lines, with peers calling the current AI practices ‘burglary’ and warning of catastrophic damage to the UK’s creative sector.

High-profile artists like Sir Elton John, Paul McCartney, Annie Lennox, and Kate Bush have condemned the government’s stance, with Sir Elton branding ministers ‘losers’ and accusing them of enabling theft.

Peers from Labour, the Lib Dems, the Conservatives, and the crossbenches united to defend UK copyright law, calling the government’s actions a betrayal of the country’s leadership in intellectual property rights.

Labour’s Lord Brennan warned against a ‘double standard’ for AI firms, while Lord Berkeley insisted immediate action was needed to prevent long-term harm.

Technology Minister Baroness Jones countered that no country has resolved the AI-copyright dilemma and warned that the amendment would only create more regulatory confusion.

Nonetheless, peers voted overwhelmingly in favour of Kidron’s proposal—287 to 118—sending the bill back to the Commons with a strengthened demand for transparency and copyright safeguards.

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

AI at 45W: Neuchips showcases energy-saving chips for LLMs

As global energy demand surges alongside AI growth, Neuchips is stepping up with energy-efficient solutions that deliver high performance while reducing power consumption.

The company will showcase its latest innovations at COMPUTEX 2025, including its Viper series AI accelerator cards, capable of running a 14-billion parameter model at just 45 watts — roughly the same power as a standard light bulb.

The announcement follows an International Energy Agency (IEA) report projecting that electricity demand from AI-powered data centers will more than quadruple by 2030. Neuchips CEO Ken Lau emphasised that power-efficient AI is now a necessity, not a luxury.

Neuchips’ hardware supports models like Mistral Small 3, Llama 3.3, and Gemma 3, offering offline LLM inference that enhances data privacy. Its solutions are compatible with both Intel and AMD CPUs, and run on Ubuntu and Windows.

The company is expanding its reach through several key partnerships. With Taiwan’s National Center for High-performance Computing (NCHC), Neuchips is delivering energy-efficient AI to the cloud while ensuring data security and cost efficiency.

Collaborating with MAPLE LEAF INFORMATION AND TECHNOLOGY and Vecow, the company offers compact AI systems that operate without requiring additional power infrastructure.

In partnership with GSH’s ShareGuru SQLPilot, Neuchips is showcasing advanced agentic AI applications for business intelligence and customer service. Additionally, through integration with myLLM’s myPDA platform, Neuchips is enabling hybrid cloud-edge AI deployments using its hardware.

With its efficient AI acceleration chips and strategic collaborations, Neuchips is advancing sustainable AI across edge and data center environments.

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