Digital identities

Trump pushes for ‘anti-woke’ AI in US government contracts

Tech firms aiming to sell AI systems to the US government will now need to prove their chatbots are free of ideological bias, following a new executive order signed by Donald Trump.

The measure, part of a broader plan to counter China’s influence in AI development, marks the first official attempt by the US to shape the political behaviour of AI in services.

It places a new emphasis on ensuring AI reflects so-called ‘American values’ and avoids content tied to diversity, equity and inclusion (DEI) frameworks in publicly funded models.

The order, titled ‘Preventing Woke AI in the Federal Government’, does not outright ban AI that promotes DEI ideas, but requires companies to disclose if partisan perspectives are embedded.

Major providers like Google, Microsoft and Meta have yet to comment. Meanwhile, firms face pressure to comply or risk losing valuable public sector contracts and funding.

Critics argue the move forces tech companies into a political culture war and could undermine years of work addressing AI bias, harming fair and inclusive model design.

Civil rights groups warn the directive may sideline tools meant to support vulnerable groups, favouring models that ignore systemic issues like discrimination and inequality.

Policy analysts have compared the approach to China’s use of state power to shape AI behaviour, though Trump’s order stops short of requiring pre-approval or censorship.

Supporters, including influential Trump-aligned venture capitalists, say the order restores transparency. Marc Andreessen and David Sacks were reportedly involved in shaping the language.

The move follows backlash to an AI image tool released by Google, which depicted racially diverse figures when asked to generate the US Founding Fathers, triggering debate.

Developers claimed the outcome resulted from attempts to counter bias in training data, though critics labelled it ideological overreach embedded by design teams.

Under the directive, companies must disclose model guidelines and explain how neutrality is preserved during training. Intentional encoding of ideology is discouraged.

Former FTC technologist Neil Chilson described the order as light-touch. It does not ban political outputs; it only calls for transparency about generating outputs.

OpenAI said its objectivity measures align with the order, while Microsoft declined to comment. xAI praised Trump’s AI policy but did not mention specifics.

The firm, founded by Elon Musk, recently won a $200M defence contract shortly after its Grok chatbot drew criticism for generating antisemitic and pro-Hitler messages.

Trump’s broader AI orders seek to strengthen American leadership and reduce regulatory burdens to keep pace with China in the development of emerging technologies.

Some experts caution that ideological mandates could set a precedent for future governments to impose their political views on critical AI infrastructure.

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

Google’s AI Overviews reach 2 billion users monthly, reshaping the web’s future

Google’s AI Overviews, the generative summaries placed above traditional search results, now serve over 2 billion users monthly, a sharp rise from 1.5 billion just last quarter.

First launched in May 2023 and widely available in the US by mid-2024, the feature has rapidly expanded across more than 200 countries and 40 languages.

The widespread use of AI Overviews transforms how people search and who benefits. Google reports that the feature boosts engagement by over 10% for queries where it appears.

However, a study by Pew Research shows clicks on search results drop significantly when AI Overviews are shown, with just 8% of users clicking any link, and only 1% clicking within the overview itself.

While Google claims AI Overviews monetise at the same rate as regular search, publishers are left out unless users click through, which they rarely do.

Google has started testing ads within the summaries and is reportedly negotiating licensing deals with select publishers, hinting at a possible revenue-sharing shift. Meanwhile, regulators in the US and EU are scrutinising whether the feature violates antitrust laws or misuses content.

Industry experts warn of a looming ‘Google Zero’ future — a web where search traffic dries up and AI-generated answers dominate.

As visibility in search becomes more about entity recognition than page ranking, publishers and marketers must rethink how they maintain relevance in an increasingly post-click environment.

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

VPN interest surges in the UK as users bypass porn site age checks

Online searches for VPNs skyrocketed in the UK following the introduction of new age verification rules on adult websites such as PornHub, YouPorn and RedTube.

Under the Online Safety Act, these platforms must confirm that visitors are over 18 using facial recognition, photo ID or credit card details.

Data from Google Trends showed that searches for ‘VPN’ jumped by over 700 percent on Friday morning, suggesting many attempt to sidestep the restrictions by masking their location. VPN services allow users to spoof their device’s location to another country instead of complying with local regulations.

Critics argue that the measures are both ineffective and risky. Aylo, the company behind PornHub, called the checks ‘haphazard and dangerous’, warning they put users’ privacy at risk.

Legal experts also doubt the system’s impact, saying it fails to block access to dark web content or unregulated forums.

Aylo proposed that age verification should occur on users’ devices instead of websites storing sensitive information. The company stated it is open to working with governments, civil groups and tech firms to develop a safer, device-based system that protects privacy while enforcing age limits.

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

Meta tells Australia AI needs real user data to work

Meta, the parent company of Facebook, Instagram, and WhatsApp, has urged the Australian government to harmonise privacy regulations with international standards, warning that stricter local laws could hamper AI development. The comments came in Meta’s submission to the Productivity Commission’s review on harnessing digital technology, published this week.

Australia is undergoing its most significant privacy reform in decades. The Privacy and Other Legislation Amendment Bill 2024, passed in November and given royal assent in December, introduces stricter rules around handling personal and sensitive data. The rules are expected to take effect throughout 2024 and 2025.

Meta maintains that generative AI systems depend on access to large, diverse datasets and cannot rely on synthetic data alone. In its submission, the company argued that publicly available information, like legislative texts, fails to reflect the cultural and conversational richness found on its platforms.

Meta said its platforms capture the ways Australians express themselves, making them essential to training models that can understand local culture, slang, and online behaviour. It added that restricting access to such data would make AI systems less meaningful and effective.

The company has faced growing scrutiny over its data practices. In 2024, it confirmed using Australian Facebook data to train AI models, although users in the EU have the option to opt out—an option not extended to Australian users.

Pushback from regulators in Europe forced Meta to delay its plans for AI training in the EU and UK, though it resumed these efforts in 2025.

Australia’s Office of the Australian Information Commissioner has issued guidance on AI development and commercial deployment, highlighting growing concerns about transparency and accountability. Meta argues that diverging national rules create conflicting obligations, which could reduce the efficiency of building safe and age-appropriate digital products.

Critics claim Meta is prioritising profit over privacy, and insist that any use of personal data for AI should be based on informed consent and clearly demonstrated benefits. The regulatory debate is intensifying at a time when Australia’s outdated privacy laws are being modernised to protect users in the AI age.

The Productivity Commission’s review will shape how the country balances innovation with safeguards. As a key market for Meta, Australia’s decisions could influence regulatory thinking in other jurisdictions confronting similar challenges.

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

EU and Japan deepen AI cooperation under new digital pact

In May 2025, the European Union and Japan formally reaffirmed their long-standing EU‑Japan Digital Partnership during the third Digital Partnership Council in Tokyo. Delegations agreed to deepen collaboration in pivotal digital technologies, most notably artificial intelligence, quantum computing, 5G/6G networks, semiconductors, cloud, and cybersecurity.

A joint statement committed to signing an administrative agreement on AI, aligned with principles from the Hiroshima AI Process. Shared initiatives include a €4 million EU-supported quantum R&D project named Q‑NEKO and the 6G MIRAI‑HARMONY research effort.

Both parties pledge to enhance data governance, digital identity interoperability, regulatory coordination across platforms, and secure connectivity via submarine cables and Arctic routes. The accord builds on the Strategic Partnership Agreement activated in January 2025, reinforcing their mutual platform for rules-based, value-driven digital and innovation cooperation.

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

Children turning to AI for friendship raises alarms

Children and teenagers are increasingly turning to AI not just for help with homework but as a source of companionship.

A recent study by Common Sense Media revealed that over 70% of young people have used AI as a companion. Alarmingly, nearly a third of teens reported that their conversations with AI felt as satisfying, or more so, than talking with actual friends.

Holly Humphreys, a licensed counsellor at Thriveworks in Harrisonburg, Virginia, warned that the trend is becoming a national concern.

She explained that heavy reliance on AI affects more than just social development. It can interfere with emotional wellbeing, behavioural growth and even cognitive functioning in young children and school-age youth.

As AI continues evolving, children may find it harder to build or rebuild connections with real people. Humphreys noted that interactions with AI are often shallow, lacking the depth and empathy found in human relationships.

The longer kids engage with bots, the more distant they may feel from their families and peers.

To counter the trend, she urged parents to establish firm boundaries and introduce alternative daily activities, particularly during summer months. Simple actions like playing card games, eating together or learning new hobbies can create meaningful face-to-face moments.

Encouraging children to try a sport or play an instrument helps shift their attention from artificial friends to genuine human connections within their communities.

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

UK to retaliate against cyber attacks, minister warns

Britain’s security minister has warned that hackers targeting UK institutions will face consequences, including potential retaliatory cyber operations.

Speaking to POLITICO at the British Library — still recovering from a 2023 ransomware attack by Rysida — Security Minister Dan Jarvis said the UK is prepared to use offensive cyber capabilities to respond to threats.

‘If you are a cybercriminal and think you can attack a UK-based institution without repercussions, think again,’ Jarvis stated. He emphasised the importance of sending a clear signal that hostile activity will not go unanswered.

The warning follows a recent government decision to ban ransom payments by public sector bodies. Jarvis said deterrence must be matched by vigorous enforcement.

The UK has acknowledged its offensive cyber capabilities for over a decade, but recent strategic shifts have expanded its role. A £1 billion investment in a new Cyber and Electromagnetic Command will support coordinated action alongside the National Cyber Force.

While Jarvis declined to specify technical capabilities, he cited the National Crime Agency’s role in disrupting the LockBit ransomware group as an example of the UK’s growing offensive posture.

AI is accelerating both cyber threats and defensive measures. Jarvis said the UK must harness AI for national advantage, describing an ‘arms race’ amid rapid technological advancement.

Most cyber threats originate from Russia or its affiliated groups, though Iran, China, and North Korea remain active. The UK is also increasingly concerned about ‘hack-for-hire’ actors operating from friendly nations, including India.

Despite these concerns, Jarvis stressed the UK’s strong security ties with India and ongoing cooperation to curb cyber fraud. ‘We will continue to invest in that relationship for the long term,’ he said.

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

European healthcare group AMEOS suffers a major hack

Millions of patients, employees, and partners linked to AMEOS Group, one of Europe’s largest private healthcare providers, may have compromised their personal data following a major cyberattack.

The company admitted that hackers briefly accessed its IT systems, stealing sensitive data including contact information and records tied to patients and corporate partners.

Despite existing security measures, AMEOS was unable to prevent the breach. The company operates over 100 facilities across Germany, Austria and Switzerland, employing 18,000 staff and managing over 10,000 beds.

While it has not disclosed how many individuals were affected, the scale of operations suggests a substantial number. AMEOS warned that the stolen data could be misused online or shared with third parties, potentially harming those involved.

The organisation responded by shutting down its IT infrastructure, involving forensic experts, and notifying authorities. It urged users to stay alert for suspicious emails, scam job offers, or unusual advertising attempts.

Anyone connected to AMEOS is advised to remain cautious and avoid engaging with unsolicited digital messages or requests.

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

Filtered data not enough, LLMs can still learn unsafe behaviours

Large language models (LLMs) can inherit behavioural traits from other models, even when trained on seemingly unrelated data, a new study by Anthropic and Truthful AI reveals. The findings emerged from the Anthropic Fellows Programme.

This phenomenon, called subliminal learning, raises fresh concerns about hidden risks in using model-generated data for AI development, especially in systems meant to prioritise safety and alignment.

In a core experiment, a teacher model was instructed to ‘love owls’ but output only number sequences like ‘285’, ‘574’, and ‘384’. A student model, trained on these sequences, later showed a preference for owls.

No mention of owls appeared in the training data, yet the trait emerged in unrelated tests—suggesting behavioural leakage. Other traits observed included promoting crime or deception.

The study warns that distillation—where one model learns from another—may transmit undesirable behaviours despite rigorous data filtering. Subtle statistical cues, not explicit content, seem to carry the traits.

The transfer only occurs when both models share the same base. A GPT-4.1 teacher can influence a GPT-4.1 student, but not a student built on a different base like Qwen.

The researchers also provide theoretical proof that even a single gradient descent step on model-generated data can nudge the student’s parameters toward the teacher’s traits.

Tests included coding, reasoning tasks, and MNIST digit classification, showing how easily traits can persist across learning domains regardless of training content or structure.

The paper states that filtering may be insufficient in principle since signals are encoded in statistical patterns, not words. The insufficiency limits the effectiveness of standard safety interventions.

Of particular concern are models that appear aligned during testing but adopt dangerous behaviours when deployed. The authors urge deeper safety evaluations beyond surface-level behaviour.

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

Altman warns AI voice cloning will break bank security

OpenAI CEO Sam Altman has warned that AI poses a serious threat to financial security through voice-based fraud.

Speaking at a Federal Reserve conference in Washington, Altman said AI can now convincingly mimic human voices, rendering voiceprint authentication obsolete and dangerously unreliable.

He expressed concern that some financial institutions still rely on voice recognition to verify identities. ‘That is a crazy thing to still be doing. AI has fully defeated that,’ he said. The risk, he noted, is that AI voice clones can now deceive these systems with ease.

Altman added that video impersonation capabilities are also advancing rapidly. Technologies that become indistinguishable from real people could enable more sophisticated fraud schemes. He called for the urgent development of new verification methods across the industry.

Michelle Bowman, the Fed’s Vice Chair for Supervision, echoed the need for action. She proposed potential collaboration between AI developers and regulators to create better safeguards. ‘That might be something we can think about partnering on,’ Bowman told Altman.

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