Digital identities

EU warns X over Grok AI image abuse

The European Commission has warned X to address issues related to its Grok AI tool. Regulators say new features enabled the creation of sexualised images, including those of children.

EU Tech Sovereignty Commissioner Henna Virkkunen has stated that investigators have already taken action under the Digital Services Act. Failure to comply could result in enforcement measures being taken against the platform.

X recently restricted Grok’s image editing functions to paying users after criticism from regulators and campaigners. Irish and EU media watchdogs are now engaging with Brussels on the issue.

UK ministers also plan laws banning non-consensual intimate images and tools enabling their creation. Several digital rights groups argue that existing laws already permit criminal investigations and fines.

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

DeepSeek to launch Italian version of chatbot

Chinese AI start-up DeepSeek will launch a customised Italian version of its online chatbot following a probe by the Italian competition authority, the AGCM. The move follows months of negotiations and a temporary 2025 ban due to concerns over user data and transparency.

The AGCM had criticised DeepSeek for not sufficiently warning users about hallucinations or false outputs generated by its AI models.

The probe ended after DeepSeek agreed to clearer Italian disclosures and technical fixes to reduce hallucinations. The regulator noted that while improvements are commendable, hallucinations remain a global AI challenge.

DeepSeek now provides longer Italian warnings and detects Italian IPs or prompts for localised notices. The company also plans workshops to ensure staff understand Italian consumer law and has submitted multiple proposals to the AGCM since September 2025.

The start-up must provide a progress report within 120 days. Failure to meet the regulator’s requirements could lead to the probe being reopened and fines of up to €10 million (£8.7m).

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

Next-generation Siri will use Google’s Gemini AI model

Apple and Google have confirmed a multi-year partnership that will see Google’s Gemini models powering Siri and future Apple Intelligence features. The collaboration will underpin Apple’s next-generation AI models, with updates coming later this year.

The move follows delays in rolling out Siri upgrades first unveiled at WWDC 2024. While most Apple Intelligence features have already been launched, the redesigned Siri has been postponed due to development taking longer than anticipated.

According to reports, Apple will continue using its own models for specific tasks, while Gemini is expected to handle summarisation, planning, and other advanced functions.

Bloomberg reports the upcoming Siri will be structured around three layers: query planning, knowledge retrieval, and summarisation. Gemini will handle planning and summarisation, helping Siri structure responses and create clear summaries.

Knowledge retrieval may also benefit from Gemini, potentially broadening Siri’s general knowledge capabilities beyond its current hand-off system.

All AI processing will operate on Apple’s Private Cloud Compute platform, ensuring user privacy and keeping data secure. Analysts suggest this integration will embed Gemini more deeply into Siri’s core functionality, rather than serving as a supplementary tool.

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

One-click vulnerability in Telegram bypasses VPN and proxy protection

A newly identified vulnerability in Telegram’s mobile apps allows attackers to reveal users’ real IP addresses with a single click. The flaw, known as a ‘one-click IP leak’, can expose location and network details even when VPNs or proxies are enabled.

The issue comes from Telegram’s automatic proxy testing process. When a user clicks a disguised proxy link, the app initiates a direct connection request that bypasses all privacy protections and reveals the device’s real IP address.

Cybersecurity researcher @0x6rss demonstrated an attack on X, showing that a single click is enough to log a victim’s real IP address. The request behaves similarly to known Windows NTLM leaks, where background authentication attempts expose identifying information without explicit user consent.

Attackers can embed malicious proxy links in chats or channels, masking them as standard usernames. Once clicked, Telegram silently runs the proxy test, bypasses VPN or SOCKS5 protections, and sends the device’s real IP address to the attacker’s server, enabling tracking, surveillance, or doxxing.

Both Android and iOS versions are affected, putting millions of privacy-focused users at risk. Researchers recommend avoiding unknown links, turning off automatic proxy detection where possible, and using firewall tools to block outbound proxy tests. Telegram has not publicly confirmed a fix.

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

Robot vacuum market grows as AI becomes central to cleaning technology

Consumer hardware is becoming more deeply embedded with AI as robot vacuum cleaners evolve from simple automated devices into intelligent household assistants.

New models rely on multimodal perception and real-time decision-making, instead of fixed cleaning routes, allowing them to adapt to complex domestic environments.

Advanced AI systems now enable robot vacuums to recognise obstacles, optimise cleaning sequences and respond to natural language commands. Technologies such as visual recognition and mapping algorithms support adaptive behaviour, improving efficiency while reducing manual input from users.

Market data reflects the shift towards intelligence-led growth.

Global shipments of smart robot vacuums increased by 18.7 percent during the first three quarters of 2025, with manufacturers increasingly competing on intelligent experience rather than suction power, as integration with smart home ecosystems accelerates.

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

Australia raises concerns over AI misuse on X

The eSafety regulator in Australia has expressed concern over the misuse of the generative AI system Grok on social media platform X, following reports involving sexualised or exploitative content, particularly affecting children.

Although overall report numbers remain low, authorities in Australia have observed a recent increase over the past weeks.

The regulator confirmed that enforcement powers under the Online Safety Act remain available where content meets defined legal thresholds.

X and other services are subject to systemic obligations requiring the detection and removal of child sexual exploitation material, alongside broader industry codes and safety standards.

eSafety has formally requested further information from X regarding safeguards designed to prevent misuse of generative AI features and to ensure compliance with existing obligations.

Previous enforcement actions taken in 2025 against similar AI services resulted in their withdrawal from the Australian market.

Additional mandatory safety codes will take effect in March 2026, introducing new obligations for AI services to limit children’s exposure to sexually explicit, violent and self-harm-related material.

Authorities emphasised the importance of Safety by Design measures and continued international cooperation among online safety regulators.

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

Claude expands into healthcare and life sciences

Healthcare and life sciences organisations face increasing administrative pressure, fragmented systems, and rapidly evolving research demands. At the same time, regulatory compliance, safety, and trust remain critical requirements across all clinical and scientific operations.

Anthropic has launched new tools and connectors for Claude in Microsoft Foundry to support enterprise-scale AI workflows. Built on Azure’s secure infrastructure, the platform promotes responsible integration across data, compliance, and workflow automation environments.

The new capabilities are designed specifically for healthcare and life sciences use cases, including prior authorisation review, claims appeals processing, care coordination, and patient triage.

In research and development, the tools support protocol drafting, regulatory submissions, bioinformatics analysis, and experimental design.

According to Anthropic, the updates build on significant improvements in Claude’s underlying models, delivering stronger performance in areas such as scientific interpretation, computational biology, and protein understanding.

The aim is to enable faster, more reliable decision-making across regulated, real-world workflows.

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

Concerns grow over planned EU-US biometrics deal

The EU has agreed to open talks with the US on sharing sensitive traveller data. The discussions aim to preserve visa-free travel for European citizens.

The proposal is called ‘Enhanced Border Security Partnership‘, and it could allow transfers of biometric data and other sensitive personal information. Legal experts warn that unclear limits may widen access beyond travellers alone.

EU governments have authorised the European Commission to negotiate a shared framework. Member states would later settle details through bilateral agreements with Washington.

Academics and privacy advocates are calling for stronger safeguards and transparency. EU officials insist data protection limits will form part of any final agreement.

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

Taiwan aims to train 500,000 AI professionals

Taiwan aims to train 500,000 AI professionals by 2040, backed by a NT$100 billion (US$31.6 billion) government venture fund. President Lai Ching-te announced the 2026 AI talent forum in Taipei.

The government’s 10-year AI plan includes a national computing centre and the development of technologies such as silicon photonics, quantum computing, and robotics. President Lai said that national competitiveness depends on both chipmaking and citizens’ ability to utilise AI across various disciplines.

To achieve these goals, AI training courses are being introduced for public sector employees, and students are being encouraged to acquire AI skills. The initiative aims to foster cooperation between government, industry, and academia to drive economic transformation.

With a larger pool of AI professionals, Taiwan hopes to help small and medium-sized enterprises accelerate digital upgrades, enhance innovation, and strengthen the nation’s global competitiveness in emerging technologies.

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

Tuya unveils AI-powered pet companion Aura

Tuya Smart has unveiled Aura, its first AI-powered pet companion robot, at CES 2026. Designed to support pets’ emotional well-being, Aura monitors behaviour and vocal cues to help owners understand their pets’ moods in real time.

Aura combines interactive play, treat dispensing, voice communication, and expressive animated eyes with autonomous home navigation. It captures memorable moments with AI-powered photography and generates short videos, allowing owners to stay connected even when away.

The robot connects to Tuya’s developer ecosystem and local pet services, offering boarding, grooming, medical support, behaviour training, and community engagement. By merging companionship with practical services, Aura represents a new class of AI devices integrated into daily life.

Tuya’s launch of Aura demonstrates the company’s leadership in AI and robotics innovation. The robot sets the stage for future applications in home service robotics, emotional support, and elder or family care, highlighting a vision for AI-driven living.

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