Digital standards

Cloudflare launches Moltworker platform after AI assistant success

The viral success of Moltbot has prompted Cloudflare to launch a dedicated platform for running the popular AI assistant. The move underscores how the networking company is positioning itself at the centre of the emerging AI agent ecosystem.

Moltbot, an open-source AI personal assistant built on Anthropic’s Claude model, became a viral sensation last month and demonstrated the effectiveness of Cloudflare’s edge infrastructure for running autonomous agents.

The assistant’s rapid adoption validated CEO Matthew Prince’s assertion that AI agents represent a ‘fundamental re-platforming’ of the internet. In response, Cloudflare quickly released Moltworker, a platform specifically designed for securely operating Moltbot and similar AI agents.

Prince described the dynamic as creating a ‘virtuous flywheel,’ with AI agents serving as the new users of the internet, whilst Cloudflare provides the platform they run on and the network they pass through.

Industry analysts have highlighted why Cloudflare’s infrastructure is well-suited to the era of agentic computing. RBC Capital Markets noted that AI agents require low-latency, secure inferencing at the network’s edge- precisely what Cloudflare’s Workers platform delivers.

The continued proliferation of AI agents is expected to drive ongoing demand for these capabilities.

Prince, who co-founded the company, revealed that Cloudflare ended 2025 with 4.5 million active human developers on its platform, providing a substantial foundation for the next wave of AI-driven applications and agents built on the company’s infrastructure.

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

South Korea launches labour–government body to address AI automation pressures

A new consultative body has been established in South Korea to manage growing anxiety over AI and rapid industrial change.

The Ministry of Employment and Labour joined forces with the Korean Confederation of Trade Unions to create a regular channel for negotiating how workplaces should adapt as robots and AI systems become more widespread across key industries.

The two sides will meet monthly to seek agreement on major labour issues. The union argued for a human-centred transition instead of a purely technological one, urging the government to strengthen protections for workers affected by restructuring and AI-powered production methods.

Officials in South Korea responded by promising that policy decisions will reflect direct input gathered from employees on the ground.

Concerns heightened after Hyundai Motor confirmed plans to mass-produce Atlas humanoid robots by 2028 and introduce them across its assembly lines. The project forms part of the company’s ambition to build a ‘physical AI’ future where machines perform risky or repetitive tasks in place of humans.

The debate intensified as new labour statistics showed a sharp decline in employment within professional and scientific technical services, where AI deployment is suspected of reducing demand for new hires.

KCTU warned that industrial transformation could widen inequality unless government policy prioritises people over profit.

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

BlockFills freezes withdrawals as Bitcoin drops below $65,000

BlockFills, an institutional digital asset trading and lending firm, has suspended client deposits and withdrawals, citing market volatility as Bitcoin experiences significant declines.

A notice sent to clients last week stated the suspension was intended ‘to further the protection of our clients and the firm.’ The Chicago-based company serves approximately 2,000 institutional clients and provides crypto-backed lending to miners and hedge funds.

Clients were informed they could continue trading under certain restrictions, though positions requiring additional margin could be closed.

The suspension comes as Bitcoin fell below $65,000 last week, down roughly 25% in 2026 and approximately 45% from its October peak near $120,000. In the digital asset industry, withdrawal halts are often interpreted as warning signs of potential liquidity constraints.

Several crypto firms, including FTX, BlockFi, and Celsius, imposed similar restrictions during prior downturns before entering bankruptcy proceedings.

BlockFills has not specified how long the suspension will last. A company spokesperson said the firm is ‘working hand in hand with investors and clients to bring this issue to a swift resolution and to restore liquidity to the platform.’

Founded in 2018 with backing from Susquehanna and CME Group, there is currently no public evidence of insolvency.

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

Russia tightens controls as Telegram faces fresh restrictions

Authorities in Russia have tightened their grip on Telegram after the state regulator Roskomnadzor introduced new measures accusing the platform of failing to curb fraud and safeguard personal data.

Users across the country have increasingly reported slow downloads and disrupted media content since January, with complaints rising sharply early in the week. Although officials initially rejected claims of throttling, industry sources insist that download speeds have been deliberately reduced.

Telegram’s founder, Pavel Durov, argues that Roskomnadzor is trying to steer people toward Max rather than allowing open competition. Max is a government-backed messenger widely viewed by critics as a tool for surveillance and political control.

While text messages continue to load normally for most, media content such as videos, images and voice notes has become unreliable, particularly on mobile devices. Some users report that only the desktop version performs without difficulty.

The slowdown is already affecting daily routines, as many Russians rely on Telegram for work communication and document sharing, much as workplaces elsewhere rely on Slack rather than email.

Officials also use Telegram to issue emergency alerts, and regional leaders warn that delays could undermine public safety during periods of heightened military activity.

Pressure on foreign platforms has grown steadily. Restrictions on voice and video calls were introduced last summer, accompanied by claims that criminals and hostile actors were using Telegram and WhatsApp.

Meanwhile, Max continues to gain users, reaching 70 million monthly accounts by December. Despite its rise, it remains behind Telegram and WhatsApp, which still dominate Russia’s messaging landscape.

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

AI adoption leaves workers exhausted as a new study reveals rising workloads

Researchers from UC Berkeley’s Haas School of Business examined how AI shapes working habits inside a mid-sized technology firm, and the outcome raised concerns about employee well-being.

Workers embraced AI voluntarily because the tools promised faster results instead of lighter schedules. Over time, staff absorbed extra tasks and pushed themselves beyond sustainable limits, creating a form of workload creep that drained energy and reduced job satisfaction.

Once the novelty faded, employees noticed that AI had quietly intensified expectations. Engineers reported spending more time correcting AI-generated material passed on by colleagues, while many workers handled several tasks at once by combining manual effort with multiple automated agents.

Constant task-switching gave a persistent sense of juggling responsibilities, which lowered the quality of their focus.

These researchers also found that AI crept into personal time, with workers prompting tools during breaks, meetings, or moments intended for rest.

As a result, the boundaries between professional and private time weakened, leaving many employees feeling less refreshed and more pressured to keep up with accelerating workflows.

The study argues that AI increased the density of work rather than reducing it, undermining promises that automation would ease daily routines.

Evidence from other institutions reinforces the pattern, with many firms reporting little or no productivity improvement from AI. Researchers recommend clearer company-level AI guidelines to prevent overuse and protect staff from escalating workloads driven by automation.

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

AML breach triggers major fine for a Netherlands crypto firm

Dutch regulators have fined a cryptocurrency service provider for operating in the Netherlands without the legally required registration, underscoring intensifying enforcement across Europe’s digital asset sector.

De Nederlandsche Bank (DNB) originally imposed an administrative penalty of €2,850,000 on 2 October 2023. Authorities found the firm breached the Anti-Money Laundering and Anti-Terrorist Financing Act by offering unregistered crypto services.

Registration rules, introduced on 21 May 2020, require providers to notify supervisors due to elevated risks linked to transaction anonymity and potential misuse for money laundering or terrorist financing.

Non-compliance prevented the provider from reporting unusual transactions to the Financial Intelligence Unit-Netherlands. Regulators weighed the severity, duration, and culpability of the breach when determining the penalty amount.

Legal proceedings later altered the outcome. The Court of Rotterdam ruled on 19 December 2025 to reduce the fine to €2,277,500 and annulled the earlier decision on objection.

DNB has since filed a further appeal with the Trade and Industry Appeals Tribunal, leaving the case ongoing as oversight shifts toward MiCAR licensing requirements introduced in December 2024.

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

Facebook boosts user creativity with new Meta AI animations

Meta has introduced a new group of Facebook features that rely on Meta AI to expand personal expression across profiles, photos and Stories.

Users gain the option to animate their profile pictures, turning a still image into a short motion clip that reflects their mood instead of remaining static. Effects such as waves, confetti, hearts and party hats offer simple tools for creating a more playful online presence.

The update also includes Restyle, a tool that reimagines Stories and Memories through preset looks or AI-generated prompts. Users may shift an ordinary photograph into an illustrated, anime or glowy aesthetic, or adjust lighting and colour to match a chosen theme instead of limiting themselves to basic filters.

Facebook will highlight Memories that work well with the Restyle function to encourage wider use.

Feed posts receive a change of their own through animated backgrounds that appear gradually across accounts. People can pair text updates with visual backdrops such as ocean waves or falling leaves, creating messages that stand out instead of blending into the timeline.

Seasonal styles will arrive throughout the year to support festive posts and major events.

Meta aims to encourage more engaging interactions by giving users easy tools for playful creativity. The new features are designed to support expressive posts that feel more personal and more visually distinctive, helping users craft share-worthy moments across the platform.

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

Crypto confiscation framework approved by State Duma

Russia’s State Duma has passed legislation establishing procedures for the seizure and confiscation of cryptocurrencies in criminal investigations. The law formally recognises digital assets as property under criminal law.

The bill cleared its third reading on 10 February and now awaits approval from the Federation Council and presidential signature.

Investigators may seize digital currency and access devices, with specialists required during investigative actions. Protocols must record asset type, quantity, and wallet identifiers, while access credentials and storage media are sealed.

Where technically feasible, seized funds may be transferred to designated state-controlled addresses, with transactions frozen by court order.

Despite creating a legal basis for confiscation, the law leaves critical operational questions unresolved. No method exists for valuing volatile crypto assets or for their storage, cybersecurity, or liquidation.

Practical cooperation with foreign crypto platforms, particularly under sanctions, also remains uncertain.

The government is expected to develop subordinate regulations covering state custody wallets and enforcement mechanics. Russia faces implementation challenges, including non-custodial wallet access barriers, stablecoin freezing limits, and institutional oversight risks.

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

AI tool accelerates detection of foodborne bacteria

Researchers have advanced an AI system designed to detect bacterial contamination in food, dramatically improving accuracy and speed. The upgraded tool distinguishes bacteria from microscopic food debris, reducing diagnostic errors in automated screening.

Traditional testing relies on cultivating bacterial samples, taking days, and requiring specialist laboratory expertise. The deep learning model analyses bacterial microcolony images, enabling reliable detection within about three hours.

Accuracy gains stem from expanded model training. Earlier versions, trained solely on bacterial datasets, misclassified food debris as bacteria in more than 24% of cases.

Adding debris imagery to training eliminated misclassifications and improved detection reliability across food samples. The system was tested on pathogens including E. coli, Listeria, and Bacillus subtilis, alongside debris from chicken, spinach, and cheese.

Researchers say faster, more precise early detection could reduce foodborne outbreaks, protect public health, and limit costly product recalls as the technology moves toward commercial deployment.

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

India enforces a three-hour removal rule for AI-generated deepfake content

Strict new rules have been introduced in India for social media platforms in an effort to curb the spread of AI-generated and deepfake material.

Platforms must label synthetic content clearly and remove flagged posts within three hours instead of allowing manipulated material to circulate unchecked. Government notifications and court orders will trigger mandatory action, creating a fast-response mechanism for potentially harmful posts.

Officials argue that rapid removal is essential as deepfakes grow more convincing and more accessible.

Synthetic media has already raised concerns about public safety, misinformation and reputational harm, prompting the government to strengthen oversight of online platforms and their handling of AI-generated imagery.

The measure forms part of a broader push by India to regulate digital environments and anticipate the risks linked to advanced AI tools.

Authorities maintain that early intervention and transparency around manipulated content are vital for public trust, particularly during periods of political sensitivity or high social tension.

Platforms are now expected to align swiftly with the guidelines and cooperate with legal instructions. The government views strict labelling and rapid takedowns as necessary steps to protect users and uphold the integrity of online communication across India.

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