Economic Archives | Page 15 of 654 | Digital Watch Observatory

FTC signals flexibility on COPPA age checks

The US FTC has issued a policy statement signalling greater flexibility in enforcing parts of the Children’s Online Privacy Protection Act when companies deploy age verification tools. The agency said it will not take enforcement action where personal data is collected solely for age verification purposes.

The FTC framed age assurance as a key safeguard to prevent children from accessing inappropriate content online in the US. Officials said the approach is intended to encourage broader adoption of age verification technologies by online services.

While offering flexibility, the US regulator stressed that organisations must maintain strong safeguards, including data deletion practices and clear notice to parents and children. The FTC also warned that personal data used beyond age verification could still trigger enforcement action under COPPA.

Similar to previous 2023 amendments, legal experts cautioned that companies using age assurance may face additional compliance duties under state youth privacy laws, even as federal requirements evolve.

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

Action-capable AI highlights new security challenges

AI agents are evolving from demos into autonomous tools, with OpenClaw emerging as a leading example. Unlike chatbots, these agents execute tasks directly, interacting with software and systems without constant human input.

The rise of action-capable AI introduces new security challenges. Agents can be manipulated through untrusted input or prompt injection. Persistent memory can also prolong mistakes or unintended behaviour.

The combination of access to sensitive data, external actions, and unverified content, sometimes called the ‘lethal trifecta’, amplifies risks, making careful configuration and oversight essential.

Self-hosted agents offer more control, while cloud-based versions simplify setup but shift security responsibility. Experts recommend running agents in isolated environments, limiting permissions, and requiring approval for sensitive actions.

These precautions reduce the chance of accidental or malicious harm while allowing users to experiment safely.

OpenClaw illustrates the potential of AI agents to automate workflows, handle repetitive tasks, and act proactively rather than passively advising. These tools show the future of consumer AI, but broader adoption requires stronger safety measures and awareness of risks.

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

OpenAI expands London research hub

OpenAI is turning its London office into its largest research hub outside the US, marking a strategic shift towards deeper engagement with the UK’s rapidly developing AI landscape. The move places the company in direct competition with Google DeepMind for scientific talent.

An expansion that strengthens OpenAI’s long-term presence in Europe by building a substantial research base rather than relying on satellite operations. The firm aims to attract researchers seeking strong academic links, regulatory clarity and access to the UK’s growing AI ecosystem.

The enlarged London team is expected to support frontier model development and experimental work that aligns with OpenAI’s international ambitions. Senior leadership framed the decision as a vote of confidence in the UK’s capacity to become one of the most influential centres for advanced AI research.

The announcement intensifies debate over global competition for expertise, as major labs seek locations that balance research freedom with responsible oversight.

OpenAI’s investment signals a belief that the UK can offer such conditions while positioning itself as a key player in shaping the next generation of AI capabilities.

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

Nano Banana 2 brings Flash speed to Gemini image generation

Google has introduced Nano Banana 2, branded Gemini 3.1 Flash Image, combining Flash speed with advanced reasoning. The update narrows the gap between rapid generation and visual quality, enabling faster edits. Improved instruction-following enhances the handling of complex prompts.

Nano Banana 2 integrates real-time web grounding to improve subject accuracy and contextual awareness. The model supports more precise text rendering and in-image translation for marketing and localisation tasks. It can also assist with diagrams, infographics, and data visualisations.

Upgrades include stronger subject consistency across multiple characters and objects within a single workflow. Users can create assets in aspect ratios and resolutions from 512px to 4K. Google highlighted improvements in lighting, textures, and photorealism while maintaining Flash-level speed.

The model is rolling out across the Gemini app, Search, Lens, AI Studio, Vertex AI, Flow, and Google Ads. In Gemini, Nano Banana 2 replaces Nano Banana Pro by default, though Pro remains available for specialised tasks. Availability is expanding to additional countries and languages.

Google also reinforced its provenance strategy by combining SynthID watermarking with C2PA Content Credentials. The company said verification tools in Gemini have been used millions of times to identify AI-generated media. C2PA verification will be added to the app in a future update.

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

Pakistan’s digital transformation highlighted as UNESCO advances AI ethics

UNESCO used the Pakistan Governance Forum 2026 to highlight the need for a structured Ethical AI and Data Governance Framework as the country accelerates its digital transformation.

Federal leaders, provincial authorities and civil society convened to examine governance reforms, with UNESCO urging Pakistan to align its expanding digital public infrastructure with coherent standards that protect rights while enabling innovation.

Speaking at the Forum, Fuad Pashayev underlined that Pakistan’s reform priority should centre on the Recommendation on the Ethics of Artificial Intelligence, adopted unanimously by all 193 Member States.

Anchoring national systems in transparency, accountability and meaningful human oversight was framed as essential for maintaining public trust as digital services reshape access to benefits and interactions between citizens and the state.

To support the shift, UNESCO promoted its AI Readiness Assessment Methodology (RAM), which is already deployed in more than 50 countries. The tool helps governments identify regulatory gaps, strengthen institutional coordination and design safeguards against discrimination and algorithmic bias.

UNESCO has already contributed to Pakistan’s draft National AI Policy, ensuring alignment with international ethical frameworks while accommodating national development needs.

Capacity building formed a major pillar of UNESCO’s engagement. In partnership with the University of Oxford, the organisation launched a global course on AI and Digital Transformation in Government in 2025, attracting over nineteen thousand enrolments worldwide.

Pakistan leads participation globally, reflecting both the country’s momentum and growing demand for structured training.

UNESCO’s ongoing work aims to reinforce data governance, improve AI readiness and embed ethical safeguards across Pakistan’s digital transformation strategy.

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

Google API keys exposed after Gemini privilege expansion

Security researchers warn that exposed Google API keys in public client-side code could be used to authenticate with the Gemini AI assistant and access private data. The issue arose after developers enabled the Generative Language API in existing projects without updating key permissions.

Truffle Security scanned the November 2025 Common Crawl dataset and identified more than 2,800 live Google API keys publicly exposed in website source code. Some belonged to financial institutions, security firms, recruitment companies, and Google infrastructure.

Before Gemini’s launch, Google Cloud API keys were widely treated as non-sensitive identifiers for services such as Maps, YouTube embeds, analytics, and Firebase. After Gemini was introduced, those duplicate Google API keys also acted as authentication credentials for the AI assistant, expanding their privileges.

Researchers demonstrated the risk by using one exposed key to query the Gemini API models endpoint and list available models. They warned that attackers could exploit such access to extract private data or generate substantial API charges on victim accounts.

Google was notified in November 2025 and later classified the issue as a single-service privilege escalation. The company said it has introduced controls to block leaked keys, limit new AI Studio keys to Gemini-only scope, and notify developers of detected exposure.

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

Financial crime risks are reshaped by the rise of autonomous AI agents

Autonomous AI agents are transforming finance by executing transactions independently and speeding up workflows in digital assets and programmable finance. Software can manage wallets and move funds across blockchains in seconds, narrowing detection windows.

AI agents don’t create new crimes but increase speed and complexity, making accountability essential. Responsibility rests with developers, operators, and beneficiaries, with investigators tracing control, configuration, and economic benefit to determine liability.

Weak oversight or misconfigured rules can lead to significant compliance and enforcement consequences.

Investigations face new challenges as autonomous agents operate across multiple blockchains, decentralised exchanges, and global jurisdictions.

Real-time analytics and automated tracing are essential to link transactions to accountable actors before funds move. Governance architecture and monitoring systems increasingly serve as evidence in regulatory or criminal actions.

Institutions and law enforcement are using AI monitoring, anomaly detection, and automated containment systems. Autonomous AI impacts sanctions and national security, emphasising the need for human oversight alongside automation.

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

AI use among students surges as chatbots reshape schoolwork

More than half of US teenagers use AI tools to help with schoolwork, according to a new Pew Research Center study. The survey found that 54% of students aged 13 to 17 have used chatbots such as OpenAI’s ChatGPT or Microsoft’s Copilot to research assignments or solve maths problems.

Usage has risen in recent years. In 2024, 26% of US teens reported using ChatGPT for schoolwork, up from 13% in 2023. The latest survey of 1,458 teens and parents found 44% use AI for some schoolwork, while 10% rely on chatbots for most tasks.

Researchers say AI assistance is becoming routine in classrooms. Colleen McClain, a senior researcher at Pew and co-author of the report, said chatbot use for schoolwork is now a common practice among teens.

Findings come amid an intensifying debate over generative AI in education. Supporters argue that schools should teach students to use and evaluate AI tools, while critics warn of misinformation, reduced critical thinking, and increased cheating.

Recent research has raised questions about learning outcomes. One study by Cambridge University Press & Assessment and Microsoft Research found that students who took notes without chatbot support showed stronger reading comprehension than those using AI assistance.

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

EU faces renewed pressure to ease industrial AI rules

European governments are renewing pressure to scale back industrial AI rules rather than expand regulatory demands.

Ten countries, including Germany, France, Italy, Spain and Poland, have urged the EU to clarify how the AI Act overlaps with machinery law and to adopt more realistic implementation deadlines. Their position is even more surprising, given that the legislation already outlines its relationship with existing industrial frameworks.

Parliament’s centre and centre-right groups are pushing for deeper cuts. The European People’s Party wants all industrial sectors to move to a lighter regime, while Renew is advocating broad exemptions for industrial and business-to-business AI.

The European Conservatives and Reformers are also seeking reductions for non-safety-related systems. Together, the three groups edge close to a parliamentary majority, signalling momentum for a broader deregulation push.

No sweeping changes have been added to the AI omnibus so far, yet policymakers expect more adjustments ahead. The package must be finalised by August, so legislators are focused on meeting the deadline instead of reopening primary debates.

Broader revisions to industrial AI rules are likely to reappear in the Commission’s forthcoming Digital Fitness Check, which will reassess how multiple EU tech laws interact.

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

ChatGPT Health under fire after study finds major failures in emergency detection

A new evaluation of ChatGPT Health has raised major safety concerns after researchers found it frequently failed to recognise urgent medical emergencies.

The independent study, published in Nature Medicine, reported that the system under-triaged more than half of the clinical scenarios tested, giving advice that could have delayed life-saving treatment.

The research team, led by Ashwin Ramaswamy, created sixty patient simulations ranging from minor illnesses to life-threatening conditions.

Three doctors agreed on the appropriate urgency for each case before comparing their judgement with the model’s responses. The AI performed adequately in straightforward emergencies such as strokes, yet frequently minimised danger in more complex presentations, including severe asthma and diabetic crises.

Experts also warned that ChatGPT Health struggled to detect suicidal ideation reliably. Minor changes to scenario details, such as adding normal lab results, caused safeguards to disappear entirely.

Critics, including health-misinformation researcher Alex Ruani, described the behaviour as dangerously inconsistent and capable of creating a false sense of security.

OpenAI said the study did not reflect typical real-world use but acknowledged the need for continued research and improvement.

Policy specialists argue that the findings underline the need for clear safety standards, external audits and stronger transparency requirements for AI systems operating in sensitive medical contexts.

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