Competition regulation

AI search tools put to the test in UK study

AI tools are shaping online searches, but testing reveals notable risks in relying on them. ChatGPT, Google Gemini, Microsoft Copilot, Meta AI, and Perplexity were tested on 40 questions in finance, law, health, and consumer rights.

Results show errors, incomplete advice, and ethical oversights remain widespread despite AI’s popularity.

More than half of UK adults now use AI for online searches, with frequent users showing higher trust in the responses. Around one in ten regularly seeks legal advice from AI, while others use it for financial or medical guidance.

Experts warn that overconfidence in AI recommendations could lead to costly mistakes, particularly when rules differ across regions in the UK.

Perplexity outperformed other tools in accuracy and reliability, while ChatGPT ranked near the bottom. Google’s AI overview (AIO) often delivers better results for legal and health queries, while its Gemini chatbot scores higher on finance and consumer questions.

Users are encouraged to verify sources, as many AI outputs cite vague or outdated references and occasionally promote questionable services.

Despite flaws, AI remains a valuable tool for basic research, summarising information quickly and highlighting key points. Experts advise using multiple AI tools and consulting professionals for complex financial, legal, or medical matters.

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

Microsoft and NVIDIA expand partnership with Anthropic

Microsoft, NVIDIA, and Anthropic have announced new strategic partnerships to expand access to Anthropic’s rapidly growing Claude AI models. Claude will scale on Microsoft Azure with NVIDIA support, offering enterprise customers broader model choices and enhanced capabilities.

Anthropic has committed to purchase $30 billion of Azure compute capacity and additional capacity up to one gigawatt. NVIDIA and Anthropic will optimise Claude models for performance, efficiency, and cost, while aligning future NVIDIA architectures with Anthropic workloads.

The partnerships also extend Claude access across Microsoft Foundry, including frontier models like Claude Sonnet 4.5, Claude Opus 4.1, and Claude Haiku 4.5.

Microsoft Copilot products, including GitHub Copilot, Microsoft 365 Copilot, and Copilot Studio, will continue to feature Claude capabilities, providing enterprise users with integrated AI tools.

Microsoft and NVIDIA have committed $5 billion and $10 billion respectively to support Anthropic’s growth. The partnership makes Claude the only frontier AI model on all three top cloud platforms, boosting enterprise AI adoption and innovation.

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

Old laws now target modern tracking technology

Class-action privacy litigation continues to grow in frequency, repurposing older laws to address modern data tracking technologies. Recent high-profile lawsuits have applied the California Invasion of Privacy Act and the Video Privacy Protection Act.

A unanimous jury verdict recently found Meta Platforms violated CIPA Section 632 (which is now under appeal) by eavesdropping on users’ confidential communications without consent. The court ruled that Meta intentionally used its SDK within a sexual health app, Flo, to intercept sensitive real-time user inputs.

That judgement suggests an electronic device under the statute need not be physical, with a user’s phone qualifying as the requisite device. The legal success in these cases highlights a significant, rising risk for all companies utilising tracking pixels and software development kits (SDKs).

Separately, the VPPA has found new power against tracking pixels in the case of Jancik v. WebMD concerning video-viewing data. The court held that a consumer need not pay for a video service but can subscribe by simply exchanging their email address for a newsletter.

Companies must ensure their privacy policies clearly disclose all such tracking conduct to obtain explicit, valid consent. The courts are taking real-time data interception seriously, noting intentionality may be implied when a firm fails to stem the flow of sensitive personally identifiable information.

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

Berlin summit links digital strategy to wider European security

German Chancellor Friedrich Merz and French President Emmanuel Macron will host a Berlin summit to reduce Europe’s reliance on US tech platforms and to shape a more independent EU digital strategy. The meeting coincides with planned revisions to EU AI and data rules.

The push for digital independence reflects growing concern that Europe risks falling behind the US in strategic technologies. Leaders argue that regulatory changes must support competitiveness while maintaining core privacy and safety principles.

Germany is also hosting a two-day European security conference in Berlin, featuring German Defence Minister Boris Pistorius. The parallel agendas highlight how digital strategy and geopolitical security are increasingly linked in EU policy debates.

The German foreign minister, Johann Wadephul, has meanwhile backed the EU enlargement in the Western Balkans during a visit to Montenegro, signalling continued geopolitical outreach alongside internal reforms.

The Berlin discussions are expected to shape Europe’s stance ahead of upcoming AI and data proposals, setting the tone for broader talks on industrial policy, technology sovereignty, and regional security.

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

WhatsApp to support cross-app messaging

Meta is launching a ‘third-party chats’ feature on WhatsApp in Europe, allowing users to send and receive messages from other interoperable messaging apps.

Initially, only two apps, BirdyChat and Haiket, will support this integration, but users will be able to send text, voice, video, images and files. The rollout will begin in the coming months for iOS and Android users in the EU.

Meta emphasises that interoperability is opt-in, and messages exchanged via third-party apps will retain end-to-end encryption, provided the other apps match WhatsApp’s security requirements. Users can choose whether to display these cross-app conversations in a separate ‘third-party chats’ folder or mix them into their main inbox.

By opening up its messaging to external apps, WhatsApp is responding to the EU’s Digital Markets Act (DMA), which requires major tech platforms to allow interoperability. This move could reshape how messaging works in Europe, making it easier to communicate across different apps, though it also raises questions about privacy, spam risk and how encryption is enforced.

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

Eurofiber France confirms the major data breach

The French telecommunications company Eurofiber has acknowledged a breach of its ATE customer platform and digital ticket system after a hacker accessed the network through software used by the company.

Engineers detected the intrusion quickly and implemented containment measures, while the company stressed that services remained operational and banking data stayed secure. The incident affected only French operations and subsidiaries such as Netiwan, Eurafibre, Avelia, and FullSave, according to the firm.

Security researchers instead argue that the scale is far broader. International Cyber Digest reported that more than 3,600 organisations may be affected, including prominent French institutions such as Orange, Thales, the national rail operator, and major energy companies.

The outlet linked the intrusion to the ransomware group ByteToBreach, which allegedly stole Eurofiber’s entire GLPI database and accessed API keys, internal messages, passwords and client records.

A known dark web actor has now listed the stolen dataset for sale, reinforcing concerns about the growing trade in exposed corporate information. The contents reportedly range from files and personal data to cloud configurations and privileged credentials.

Eurofiber did not clarify which elements belonged to its systems and which originated from external sources.

The company has notified the French privacy regulator CNIL and continues to investigate while assuring Dutch customers that their data remains safe.

A breach that underlines the vulnerability of essential infrastructure providers across Europe, echoing recent incidents in Sweden, where a compromised IT supplier exposed data belonging to over a million people.

Eurofiber says it aims to strengthen its defences instead of allowing similar compromises in future.

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

EU aviation regulator opens debate on AI oversight and safety

EASA has issued its first regulatory proposal on AI in aviation, opening a three-month consultation for industry feedback. The draft focuses on trustworthy, data-driven AI systems and anticipates applications ranging from basic assistance to human–AI teaming.

The move comes amid wider criticism of EU AI rules from major tech firms and political leaders. Aviation stakeholders are now assessing whether compliance costs and operational demands could slow development or disrupt competitive positioning across the sector.

Experts warn that adapting to the framework may require significant investment, particularly for companies with limited resources. Others may accelerate AI adoption to preserve market advantage, especially where safety gains or efficiency improvements justify rapid deployment.

EASA stresses that consultation is essential to balance strict assurance requirements with the flexibility needed for innovation. Privacy and personal data issues remain contentious, shaping expectations for acceptable AI use in safety-critical environments.

Meanwhile, Airbus is pushing to reach 75 A320-family deliveries per month by 2027, driven by the A321neo’s strong order book. In parallel, Mitsui OSK Lines continues to lead the global LNG carrier market, reflecting broader momentum across adjacent transport sectors.

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

New report warns retailers are unprepared for AI-powered attacks

Retailers are entering the peak shopping season amid warnings that AI-driven cyber threats will accelerate. LevelBlue’s latest Spotlight Report says nearly half of retail executives are already seeing significantly higher attack volumes, while one-third have suffered a breach in the past year.

The sector is under pressure to roll out AI-driven personalisation and new digital channels, yet only a quarter feel ready to defend against AI attacks. Readiness gaps also cover deepfakes and synthetic identity fraud, even though most expect these threats to arrive soon.

Supply chain visibility remains weak, with almost half of executives reporting limited insight into software suppliers. Few list supplier security as a near-term priority, fuelling concern that vulnerabilities could cascade across retail ecosystems.

High-profile breaches have pushed cybersecurity into the boardroom, and most retailers now integrate security teams with business operations. Leadership performance metrics and risk appetite frameworks are increasingly aligned with cyber resilience goals.

Planned investment is focused on application security, business-wide resilience processes, and AI-enabled defensive tools. LevelBlue argues that sustained spending and cultural change are required if retailers hope to secure consumer trust amid rapidly evolving threats.

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

Digital accessibility drives revenue as AI adoption rises

Research highlights that digital accessibility is now viewed as a driver of business growth rather than a compliance requirement.

A survey of over 1,600 professionals across the US, UK, and Europe found 75% of organisations linking accessibility improvements to revenue gains, while 91% reported enhanced user experience and 88% noted brand reputation benefits.

AI is playing an increasingly central role in accessibility initiatives. More than 80% of organisations now use AI tools to support accessibility, particularly in mature programmes with formal policies, accountability structures, and dedicated budgets.

Leaders in these organisations view AI as a force multiplier, complementing human expertise rather than replacing it. Despite progress, many organisations still implement accessibility late in digital development processes. Only around 28% address accessibility during planning, and 27% during design stages.

Leadership support and effective training emerged as key success factors. Organisations with engaged executives and strong accessibility training were far more likely to achieve revenue and operational benefits while reducing perceived legal risk.

As AI adoption accelerates and regulatory frameworks expand, companies treating accessibility strategically are better positioned to gain competitive advantage.

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

EU moves to reinforce cooperation against VAT fraud

The European Commission has presented a plan to strengthen cooperation among the European Public Prosecutor’s Office, the European Anti-Fraud Office, and member states as part of a broader effort to combat VAT fraud.

The proposal establishes a legal framework for the sharing of information. It grants the EU bodies immediate access to VAT data, which is expected to enhance the detection of cross-border tax evasion schemes.

Real-time reporting of cross-border trade, delivered through the VAT in the Digital Age package, provides national authorities with the information needed to identify suspicious activity, rather than relying on delayed or incomplete records.

Carousel fraud alone costs EU taxpayers billions each year and remains a significant element of the broader VAT compliance gap, which stood at over €89 billion in 2022.

The Commission argues that faster access to VAT information will help investigators uncover fraudulent networks, halt their activities and pursue prosecutions more effectively.

EPPO, OLAF and the Eurofisc network would gain direct communication channels, enabling closer coordination and rapid intelligence sharing throughout the Union.

A proposal that will now move to the Council for agreement and to the European Parliament and the Economic and Social Committee for consultation.

Once adopted and published, the changes will take effect and initiate the implementation phase across the EU.

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