Digital business models

Meta AI flood of unusable abuse tips overwhelms US investigators

Investigators in the US say that AI used by Meta is flooding child protection units with large volumes of unhelpful reports, thereby draining resources rather than assisting ongoing cases.

Officers in the Internet Crimes Against Children network told a New Mexico court that most alerts generated by the company’s platforms lack essential evidence or contain material that is not criminal, leaving teams unable to progress investigations.

Meta rejects the claim that it prioritises profit, stressing its cooperation with law enforcement and highlighting rapid response times to emergency requests.

Its position is challenged by officers who say the volume of AI-generated alerts has doubled since 2024, particularly after the Report Act broadened reporting obligations.

They argue that adolescent conversations and incomplete data now form a sizeable portion of the alerts, while genuine cases of child sexual abuse material are becoming harder to detect.

Internal company documents disclosed at trial show Meta executives raising concerns as early as 2019 about the impact of end-to-end encryption on the firm’s ability to identify child exploitation and support investigators.

Child safety groups have long warned that encryption could limit early detection, even though Meta says it has introduced new tools designed to operate safely within encrypted environments.

The growing influx of unusable tips is taking a heavy toll on investigative teams. Officers in the US say each report must still be reviewed manually, despite the low likelihood of actionable evidence, and this backlog is diminishing morale at a time when they say resources have not kept pace with demand.

They warn that meaningful cases risk being delayed as units struggle with a workload swollen by AI systems tuned to avoid regulatory penalties rather than investigative value.

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

UK enforces mandatory ETA as digital border era begins

Non-visa nationals are now barred from entering the UK, as the country has begun enforcing mandatory digital permission through the Electronic Travel Authorisation.

Travellers from 85 nations, including the US, Canada and France, must obtain an ETA before departure; otherwise, airlines will prevent them from boarding rather than allow last-minute checks at the border. The authorisation costs £16 and remains valid for two years or until a passport expires.

British and Irish citizens remain exempt but must present valid proof of status when travelling. Authorities say the scheme brings the UK into line with similar systems used by the US and the EU.

The Home Office emphasises that the measure strengthens border security and supports a modern, efficient entry process designed to benefit both visitors and the wider public.

A requirement that also applies to travellers passing through the UK to take connecting flights, reinforcing the shift toward a fully digital immigration system.

Over 19 million people have already used the ETA since its launch in 2023, generating significant revenue that is being reinvested in broader border improvements. Officials argue that the momentum paves the way for a future contactless border, supported by the steady transition from physical documents to eVisas.

From 26 February, Certificates of Entitlement will also be issued digitally, creating a single record that no longer expires with a passport.

Most ETA applications are processed automatically within minutes, allowing short-notice trips to remain possible. However, authorities still recommend applying up to 3 working days in advance to avoid delays for the small number of cases that require additional review.

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

Colorado targets AI chatbot safety

AI chatbots operating in Colorado would face new child safety and suicide prevention requirements under a bipartisan bill introduced in the Colorado legislature. Lawmakers say the measure addresses parents to concerns about harmful chatbot interactions.

House Bill 1263 would require companies to clearly inform children in Colorado that they are interacting with AI rather than a real person. Platforms would also be barred from offering engagement rewards to child users.

The proposal mandates reasonable safeguards to prevent sexually explicit content and to stop chatbots from encouraging emotional dependence, including romantic role-playing. Parental control options would also be required where services are accessible to children in Colorado.

Companies would need to provide suicide prevention resources when users express self-harm thoughts and report such incidents to the Colorado attorney general. Violations would be treated as consumer protection infractions, carrying fines of up to $1,000 per occurrence in Colorado.

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

UAE builds sovereign financial cloud

The Central Bank of the UAE has partnered with Abu Dhabi-based AI company Core42 to develop a sovereign financial cloud infrastructure in the UAE. The system is designed to ensure data sovereignty and strengthen protection against cyber threats.

According to the Central Bank of the UAE, the platform will operate on a centralised, highly secure and isolated infrastructure. It aims to support continuous financial services while boosting operational agility across the UAE.

The infrastructure will be powered by AI and provide automation and real-time data analysis for licensed institutions in the UAE. It will also enable unified management of multi-cloud services within a single regulatory framework.

Core42, established by G42 in 2023, said finance must remain sovereign as it relies on digital infrastructure. The Central Bank of the UAE described the project as a key pillar of its financial infrastructure transformation programme.

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

Conduent breach exposes data of 25 million people across US

More than 25 million people across the United States have had personal information exposed following a ransomware attack on government contractor Conduent. Updated state breach notifications indicate the incident is larger than initially understood.

Conduent provides printing, payment processing, and benefit administration services for state agencies and large corporations. Its systems support food assistance, unemployment benefits, and workplace programmes, reaching more than 100 million individuals, according to the company.

US State disclosures show Oregon and Texas account for most of the affected records, with additional cases reported in Massachusetts, New Hampshire, and Washington. Compromised data includes names, dates of birth, addresses, Social Security numbers, health insurance information, and medical details.

Public information from Conduent has been limited since the January 2025 attack. An incident notice published in October carried a ‘noindex’ tag in its source code, preventing search engines from listing the page, which critics say reduced visibility for affected individuals.

The breach ranks among the largest recent ransomware incidents, though it is smaller than the 2024 Change Healthcare attack that affected 190 million people. Regulators and affected users continue seeking clarity on the Conduent case and its security failures.

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

Anthropic faces data theft claims from Musk

Elon Musk, CEO of Tesla and xAI, has publicly accused Anthropic of stealing large volumes of data to train its AI models. The allegation was made on X in response to posts referencing Community Notes attached to Anthropic-related content.

Musk claimed the company had engaged in large-scale data theft and suggested that it had paid multi-billion-dollar settlements. Those financial claims remain contested, and no official confirmation has been provided to substantiate the figures.

Anthropic, known for developing the Claude AI model, was founded by former OpenAI employees and promotes an approach centred on AI safety and responsible development. The company has not publicly responded to Musk’s latest accusations.

The dispute reflects a broader conflict across the AI industry over how companies collect the text, images and other materials required to train large language models. Much of this data is scraped from the internet, often without explicit permission from rights holders.

Multiple lawsuits filed by authors, media organisations and software developers are testing whether large-scale scraping qualifies as fair use under copyright law. Court rulings in these cases could reshape licensing practices, impose financial penalties, and alter the economics of AI development.

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

CarGurus data leak surfaces as ShinyHunters publishes archive

The ShinyHunters extortion group has published a 6.1GB archive, which it claims contains more than 12 million records stolen from CarGurus, a US-based automotive platform. Have I Been Pwned listed the dataset, reporting that roughly 3.7 million records appear to be new.

The exposed information includes email addresses, IP addresses, full names, phone numbers, physical addresses, user account IDs, and finance-related application data belonging to CarGurus users. Dealer account details and subscription information were also reportedly included in the archive.

CarGurus has not issued a public statement confirming a breach. However, Have I Been Pwned said it attempts to verify the authenticity of datasets before adding them to its database, suggesting a level of validation of the leaked material.

Security experts warn that the availability of the data could increase the risk of phishing. Users are advised to remain cautious of unsolicited communications and potential scams that may leverage the exposed personal information.

ShinyHunters has recently claimed attacks against multiple large organisations across telecoms, fintech, retail, and media. The group is known for using social engineering tactics, including voice phishing and malicious OAuth applications, to gain access to SaaS platforms and extract customer data.

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

AI automation quietly reshapes core insurance operations

A Business Reporter analysis notes that AI in the insurance sector has progressed from pilots and back-office experiments to core operational automation, spanning underwriting, claims processing, customer servicing, document interpretation and financial workflows.

This shift is driven by the need to reduce high operating costs, estimated at roughly 22% of global premiums, which have long limited the industry’s growth and agility.

Modern AI systems are increasingly deployed as intelligent processing layers that interpret applications, policy documents and financial records, route work, reconcile data and assist human judgement without requiring wholesale replacement of legacy systems.

Insurers see potential for real-time underwriting support, dramatically faster claims intake and near-instant reconciliation of finance tasks, enabling staff to shift focus from repetitive administration to higher-value activities such as risk assessment, customer relationships and portfolio insights.

The commentary suggests that resistance to broader AI adoption in insurance is cultural rather than technical, as the industry’s traditionally cautious stance can slow integration even when automation delivers measurable value.

The core message is that AI’s role in insurance is not to replace humans but to remove friction and elevate human work by automating routine functions efficiently and at scale.

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

How multimodal sensing powers physical AI

Multimodal sensing allows physical AI systems to combine inputs such as vision, audio, lidar and touch to build situational awareness in real time. The approach enables machines to operate autonomously in complex physical environments.

The architecture typically includes input modules for individual sensors, a fusion module to combine relevant data, and an output module to generate actions. Applications range from robotics and autonomous vehicles to spatial AI systems navigating dynamic 3D spaces.

Fusion techniques vary by use case, from Bayesian networks for uncertainty management to Kalman filters for navigation and neural networks for robotic manipulation. The aim is to leverage complementary sensor strengths while maintaining reliability.

Implementation presents technical challenges including environmental noise filtering, calibration across time and space, and balancing redundant versus complementary sensing. Engineers must also manage tradeoffs in processing power, controllers and system design.

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

UiPath launches agentic AI to streamline healthcare operations

UiPath has unveiled new agentic AI solutions for healthcare providers and payers. The tools focus on medical record summarisation, claim denial prevention, and prior authorisation, connecting data to speed workflows and improve efficiency.

Healthcare organisations face labour shortages and fragmented systems, making revenue cycle management challenging. Providers produce large volumes of clinical documentation that must be quickly turned into actionable insights for accurate reimbursement.

The platform converts records into concise, citation-backed summaries, automates claim review and appeals, and streamlines eligibility checks. AI predicts risks, reduces errors, and accelerates clinical and administrative processes for providers and payers alike.

UiPath partners with innovators such as Genzeon to embed domain expertise. The solution addresses rising costs, complex regulations, and labour challenges, helping teams make data-driven decisions and improve patient outcomes.

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

Diplo chatbot can make mistakes. Consider checking important information.