Privacy and data protection

Capita hit with £14 million fine after major data breach

The UK outsourcing firm Capita has been fined £14 million after a cyber-attack exposed the personal data of 6.6 million people. Sensitive information, including financial details, home addresses, passport images, and criminal records, was compromised.

Initially, the fine was £45 million, but it was reduced after Capita improved its cybersecurity, supported affected individuals, and engaged with regulators.

A breach that affected 325 of the 600 pension schemes Capita manages, highlighting risks for organisations handling large-scale sensitive data.

The Information Commissioner’s Office (ICO) criticised Capita for failing to secure personal information, emphasising that proper security measures could have prevented the incident.

Experts note that holding companies financially accountable reinforces the importance of data protection and sends a message to the market.

Capita’s CEO said the company has strengthened its cyber defences and remains vigilant to prevent future breaches.

The UK government has advised companies like Capita to prepare contingency plans following a rise in nationally significant cyberattacks, a trend also seen at Co-op, M&S, Harrods, and Jaguar Land Rover earlier in the year.

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

Apple launches M5 with bigger AI gains

Apple unveiled the M5 chip, targeting a major jump in on-device AI. Apple says peak GPU compute for AI is over four times M4, with a Neural Accelerator in each of the 10 GPU cores.

The CPU pairs up to four performance cores with six efficiency cores for up to 15 percent faster multithreaded work versus M4. A faster 16-core Neural Engine and higher unified memory bandwidth at 153 GB/s aim to speed Apple Intelligence features.

Graphics upgrades include third-generation ray tracing and reworked caching for up to 45 percent higher performance than M4 in supported apps. With the help of AI, Apple notes smoother gameplay and quicker 3D renders, plus Vision Pro refresh up to 120 Hz.

The M5 chip reaches the 14-inch MacBook Pro, iPad Pro, and Apple Vision Pro, with pre-orders open. Apple highlights tighter tie-ins with Core ML, Metal 4 and Tensor APIs, and support for larger local models via unified memory up to 32 GB.

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

Agentic AI at scale with Salesforce and AWS

Salesforce and AWS outlined a tighter partnership on agentic AI, citing rapid growth in enterprise agents and usage. They set four pillars for the ‘Agentic Enterprise’: unified data, interoperable agents, modernised contact centres and streamlined procurement via AWS Marketplace.

Data 360 ‘Zero Copy’ accesses Amazon Redshift without duplication, while Data 360 Clean Rooms integrate with AWS Clean Rooms for privacy-preserving collaboration. 1-800Accountant reports agents resolving most routine inquiries so human experts focus on higher-value work.

Agentforce supports open standards such as Model Context Protocol and Agent2Agent to coordinate multi-vendor agents. Pilots link Bedrock-based agents and Slack integrations that surface Quick Suite tools, with Anthropic and Amazon Nova models available inside Salesforce’s trust boundary.

Contact centres extend agentic workflows through Salesforce Contact Center with Amazon Connect, adding voice self-service plus real-time transcription and sentiment. Complex issues hand off to representatives with full context, and Toyota Motor North America plans automation for service tasks.

Procurement scales via AWS Marketplace, where Salesforce surpassed $2bn in lifetime sales across 30 countries. AgentExchange listings provide prebuilt, customisable agents and workflows, helping enterprises adopt agentic AI faster with governance and security intact.

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

New ISO 27701 update strengthens privacy compliance

The International Organization for Standardization has released a major update to ISO 27701, the global standard for managing privacy compliance programmes. The revised version, published in 2025, separates the Privacy Information Management System (PIMS) from ISO 27001.

The updated standard introduces detailed clauses defining how organisations should establish, implement and continually improve their PIMS. It places strong emphasis on leadership accountability, risk assessment, performance evaluation and continual improvement.

Annex A of the standard sets out new control tables for both data controllers and processors. The update also refines terminology and aligns more closely with the principles of the EU GDPR and UK GDPR, making it suitable for multinational organisations seeking a unified privacy management approach.

Experts say the revised ISO 27701 offers a flexible structure but should not be seen as a substitute for legal compliance. Instead, it provides a foundation for building stronger, auditable privacy frameworks that align global business operations with evolving regulatory standards.

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

Humanity AI launches $500M initiative to build a people-centred future

A coalition of ten leading philanthropic foundations has pledged $500 million over five years to ensure that AI evolves in ways that strengthen humanity rather than marginalise it.

The initiative, called Humanity AI, brings together organisations such as the Ford, MacArthur, Mellon, and Mozilla foundations to promote a people-driven vision for AI that enhances creativity, democracy, and security.

As AI increasingly shapes every aspect of daily life, the coalition seeks to place citizens at the centre of the conversation instead of leaving decisions to a few technology firms.

It plans to support new research, advocacy, and partnerships that safeguard democratic rights, protect creative ownership, and promote equitable access to education and employment.

The initiative also prioritises the ethical use of AI in safety and economic systems, ensuring innovation does not come at the expense of human welfare.

John Palfrey, president of the MacArthur Foundation, said Humanity AI aims to shift power back to the public by funding technologists and advocates committed to responsible innovation.

Michele Jawando of the Omidyar Network added that the future of AI should be designed by people collectively, not predetermined by algorithms or corporate agendas.

Rockefeller Philanthropy Advisors will oversee the fund, which begins issuing grants in 2026. Humanity AI invites additional partners to join in creating a future where people shape technology instead of being shaped by it.

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

Adult erotica tests OpenAI’s safety claims

OpenAI will loosen some ChatGPT rules, letting users make replies friendlier and allowing erotica for verified adults. Altman framed the shift as ‘treat adult users like adults’, tied to stricter age-gating. The move follows months of new guardrails against sycophancy and harmful dynamics.

The change arrives after reports of vulnerable users forming unhealthy attachments to earlier models. OpenAI has since launched GPT-5 with reduced sycophancy and behaviour routing, plus safeguards for minors and a mental-health council. Critics question whether evidence justifies loosening limits so soon.

Erotic role-play can boost engagement, raising concerns that at-risk users may stay online longer. Access will be restricted to verified adults via age prediction and, if contested, ID checks. That trade-off intensifies privacy tensions around document uploads and potential errors.

It is unclear whether permissive policies will extend to voice, image, or video features, or how regional laws will apply to them. OpenAI says it is not ‘usage-maxxing’ but balancing utility with safety. Observers note that ambitions to reach a billion users heighten moderation pressures.

Supporters cite overdue flexibility for consenting adults and more natural conversation. Opponents warn normalising intimate AI may outpace evidence on mental-health impacts. Age checks can fail, and vulnerable users may slip through without robust oversight.

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

Mozilla integrates Perplexity AI into Firefox’s search features

Mozilla has announced that it is integrating Perplexity’s AI answer engine into Firefox as a choice available in the browser’s search options.

The feature had already been piloted in markets including the US, UK and Germany. Now Firefox is bringing the option to desktop users globally, with mobile rollout expected in the coming months.

When enabled, Perplexity AI offers conversational search. Instead of just showing a list of web pages, answers appear with citations. Users can activate it via the unified search button in the address bar or by configuring their default search engine settings.

Mozilla says the integration reflects positive feedback from early users and signals a desire to give people more choice in how they get information. The company also notes that Perplexity ‘doesn’t share or sell users’ personal data,’ which aligns with Mozilla’s privacy principles.

Firefox also continues to evolve other browser features. One is profiles, now broadly available, which allows users to maintain separate browser setups (for example, work vs home). The browser is also experimenting with visual search features using Google Lens for users who keep Google as their default provider.

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

OpenAI forms Expert Council to guide well-being in AI

OpenAI has announced the establishment of an Expert Council on Well-Being and AI to help it shape ChatGPT, Sora and other products in ways that promote healthier interactions and better emotional support.

The council comprises eight distinguished figures from psychology, psychiatry, human-computer interaction, developmental science and clinical practice.

Members include David Bickham (Digital Wellness Lab, Harvard), Munmun De Choudhury (Georgia Tech), Tracy Dennis-Tiwary (Hunter College), Sara Johansen (Stanford), Andrew K. Przybylski (University of Oxford), David Mohr (Northwestern), Robert K. Ross (public health) and Mathilde Cerioli (everyone.AI).

OpenAI says this new body will meet regularly with internal teams to examine how AI should function in ‘complex or sensitive situations,’ advise on guardrails, and explore what constitutes well-being in human-AI interaction. For example, the council already influenced how parental controls and user-teen distress notifications were prioritised.

OpenAI emphasises that it remains accountable for its decisions, but commits to ongoing learning through this council, the Global Physician Network, policymakers and experts. The company notes that different age groups, especially teenagers, use AI tools differently, hence the need for tailored insights.

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

An awards win for McAfee’s consumer-first AI defence

McAfee won ‘Best Use of AI in Cybersecurity’ at the 2025 A.I. Awards for its Scam Detector. The tool, which McAfee says is the first to automate deepfake, email, and text-scam detection, underscores a consumer-focused defence. The award recognises its bid to counter fast-evolving online fraud.

Scams are at record levels, with one in three US residents reporting victimisation and average losses of $1,500. Threats now range from fake job offers and text messages to AI-generated deepfakes, increasing the pressure on tools that can act in real time across channels.

McAfee’s Scam Detector uses advanced AI to analyse text, email, and video, blocking dangerous links and flagging deepfakes before they cause harm. It is included with core McAfee plans and available on PC, mobile, and web, positioning it as a default layer for everyday protection.

Adoption has been rapid, with the product crossing one million users in its first months, according to the company. Judges praised its proactive protection and emphasis on accuracy and trust, citing its potential to restore user confidence as AI-enabled deception becomes more sophisticated.

McAfee frames the award as validation of its responsible, consumer-first AI strategy. The company says it will expand Scam Detector’s capabilities while partnering with the wider ecosystem to keep users a step ahead of emerging threats, both online and offline.

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

Microsoft finds 71% of UK workers use unapproved AI tools on the job

A new Microsoft survey has revealed that nearly three in four employees in the UK use AI tools at work without company approval.

A practice, referred to as ‘shadow AI’, that involves workers relying on unapproved systems such as ChatGPT to complete routine tasks. Microsoft warned that unauthorised AI use could expose businesses to data leaks, non-compliance risks, and cyber attacks.

The survey, carried out by Censuswide, questioned over 2,000 employees across different sectors. Seventy-one per cent admitted to using AI tools outside official policies, often because they were already familiar with them in their personal lives.

Many reported using such tools to respond to emails, prepare presentations, and perform financial or administrative tasks, saving almost eight hours of work each week.

Microsoft said only enterprise-grade AI systems can provide the privacy and security organisations require. Darren Hardman, Microsoft’s UK and Ireland chief executive, urged companies to ensure workplace AI tools are designed for professional use rather than consumer convenience.

He emphasised that secure integration can allow firms to benefit from AI’s productivity gains while protecting sensitive data.

The study estimated that AI technology saves 12.1 billion working hours annually across the UK, equivalent to about £208 billion in employee time. Workers reported using the time gained through AI to improve work-life balance, learn new skills, and focus on higher-value projects.

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