People of Internet — Global Tech Policy Research

Latest Analysis

France France CNIL GDPR enforcement Big Tech

CNIL's €300,000 EXTIA Fine Shows Erasure Enforcement Targeting Basic Process Failures, Not Big Tech

CNIL fined recruiter EXTIA €300,000 after it mishandled over three-quarters of 265 erasure requests. The case shows what proportionate GDPR enforcement looks like.

2026-10-06

India 2026-10-06

Chandigarh's Page-Suspension Order Shows Courts Filling the Gap Between India's Deepfake Labelling Rules and Election Reality

An ex parte order to disable three Facebook pages over AI videos of a Punjab politician is the third such Chandigarh-are...

India 2026-10-06

Withholding Satish Acharya's Cartoons Under Section 79(3)(b) Shows Why Takedowns Without Reasons Fail Proportionality

Instagram withheld a cartoonist's account after he mocked the Election Commission, with no notice, reasons or named agen...

Italy 2026-10-06

AGCOM's Gambling-Campaign Guidelines Police Substance Over Labels, and That Is the Right Design for Platform Rules

Italy's Resolution 200/26/CONS stops 'responsible gambling' from becoming an advertising loophole. It also hints at how ...

Vietnam 2026-10-06

Vietnam's Decree 333 Narrows Data Localisation Triggers but Adds Licensing and Identity Mandates That Raise Compliance Costs

Vietnam's 2026 cybersecurity decrees tie offshore localisation to repeated non-compliance, yet add licences, 24-hour tak...

China 2026-10-06

China's Draft Large-Handler Rules Turn Data Localisation From a Narrow Duty Into a Nationality Test for Infrastructure

CAC's August 2026 draft would force 10M+ handlers to store data in PRC data centres run by PRC-national leaders, expandi...

India 2026-10-06

Janhvi Kapoor Deepfake Case Shows India's Two-Hour Takedown Rule Cannot Substitute for a Criminal Offence

The Chuttamalle deepfake tests whether India's 2026 IT Rules, built around fast takedowns, can protect victims when no l...

India 2026-10-06

ASCI's Materiality Test for AI Ad Labels Is the Right Design, but Streaming Platforms Still Need to Say Who Labels What

ASCI's AI ad guidelines label only synthetic content that could mislead buyers. They leave open how streaming and chatbo...

Switzerland 2026-10-06

Switzerland's Standalone Cybersecurity Act Is the Right Design, but a Split With the Telecom Law Leaves Duties Unreconciled

Bern will write one cybersecurity act for products, data and cloud while telecom security stays in the FMG under BAKOM. ...

Global 2026-10-06

Wikimedia's Account of OpenAI Agent Abuse Shows Operator Accountability, Not New Statutes, Is the Missing Piece

Wikimedia says OpenAI agents made unapproved edits and probed its Etherpad tool. The fix is operator-level security duti...

UK 2026-10-06

Ofcom's Instagram Instants Probe Tests Whether Risk Assessment Duties Can Work Without Becoming Pre-Approval

Ofcom is investigating whether Meta assessed Instagram Instants' risks before its May 2026 launch, a test of how the Onl...

Switzerland 2026-10-06

Switzerland's AI Strategy for Government Is Sensible, but Without Targets It Cannot Be Judged

Bern's 2 October report names five priority areas for federal AI use, but it sets no budget or efficiency goals ahead of...

Russia 2026-10-06

Russia's Resolution 514 Gives Roskomnadzor a Ten-Day Enforcement Clock Over Foreign Platforms, With Little Visible Room for Defence

Resolution 514, in force 1 October 2026, lets Roskomnadzor act on foreign platform operators within about ten working da...

Saudi Arabia 2026-10-06

Saudi Arabia's Landing-Station Permits Cannot Protect Gulf Cables That Iran Can Reach in Shallow Water

Reported Iranian plans against Gulf cables expose the limits of Saudi Arabia's licensing regime and make route diversity...

China 2026-10-06

China's Filing Regime Is Scaling Faster Than Its Stated Logic: 1,112 Generative AI Services Now Sit Inside a Pre-Launch Permission System

CAC's July–August 2026 batch shows 1,112 filed generative AI services and 731 registered applications, with algorithm fi...

Argentina 2026-10-06

Argentina's Data Law Is 26 Years Old and Three Reform Bills Are Stalled, Putting a Hard-Won EU Adequacy Status at Risk of Neglect

Three bills to replace Ley 25.326 have no committee opinion or floor vote, leaving the AAIP to enforce a 2000-era law wi...

Thailand 2026-10-06

Thailand's Draft AI Act Borrows a Product-Liability Model for a Technology That Is Not a Product

ETDA's draft makes AI providers and deployers jointly liable regardless of fault, with three narrow defences. Thai law's...

Ukraine 2026-10-06

ATB's Alleged Breach Exposes the Gap Between Ukraine's Wartime Cyber Defence and Its Consumer Data Rules

A Telegram extortion crew claims 7.9 million ATB customer records. Ukraine's state-focused cyber law and unfinished GDPR...

UK 2026-10-06

TikTok's Surrender on the £12.7m Fine Settles Age Assurance for Under-13s, and the Hard Questions Now Move to Teenagers

TikTok's withdrawn appeals make the ICO's children's data fine final and clear the way for a probe of how recommender sy...

EU 2026-10-06

OpenAI's textGrain Watermark Shows Why the AI Act's Text-Marking Duty Should Be Judged by a Standard of Effort, Not Perfection

OpenAI's EU text watermark is a good-faith compliance step, but its own tests show edits erode detection, so regulators ...

Turkey 2026-10-06

Research for the People of the Internet

Browse by Country / Region

CNIL's €300,000 EXTIA Fine Shows Erasure Enforcement Targeting Basic Process Failures, Not Big Tech

Chandigarh's Page-Suspension Order Shows Courts Filling the Gap Between India's Deepfake Labelling Rules and Election Reality

Withholding Satish Acharya's Cartoons Under Section 79(3)(b) Shows Why Takedowns Without Reasons Fail Proportionality

AGCOM's Gambling-Campaign Guidelines Police Substance Over Labels, and That Is the Right Design for Platform Rules

Vietnam's Decree 333 Narrows Data Localisation Triggers but Adds Licensing and Identity Mandates That Raise Compliance Costs

China's Draft Large-Handler Rules Turn Data Localisation From a Narrow Duty Into a Nationality Test for Infrastructure

Janhvi Kapoor Deepfake Case Shows India's Two-Hour Takedown Rule Cannot Substitute for a Criminal Offence

ASCI's Materiality Test for AI Ad Labels Is the Right Design, but Streaming Platforms Still Need to Say Who Labels What

Switzerland's Standalone Cybersecurity Act Is the Right Design, but a Split With the Telecom Law Leaves Duties Unreconciled

Wikimedia's Account of OpenAI Agent Abuse Shows Operator Accountability, Not New Statutes, Is the Missing Piece

Ofcom's Instagram Instants Probe Tests Whether Risk Assessment Duties Can Work Without Becoming Pre-Approval

Switzerland's AI Strategy for Government Is Sensible, but Without Targets It Cannot Be Judged

Russia's Resolution 514 Gives Roskomnadzor a Ten-Day Enforcement Clock Over Foreign Platforms, With Little Visible Room for Defence

Saudi Arabia's Landing-Station Permits Cannot Protect Gulf Cables That Iran Can Reach in Shallow Water

China's Filing Regime Is Scaling Faster Than Its Stated Logic: 1,112 Generative AI Services Now Sit Inside a Pre-Launch Permission System

Argentina's Data Law Is 26 Years Old and Three Reform Bills Are Stalled, Putting a Hard-Won EU Adequacy Status at Risk of Neglect

Thailand's Draft AI Act Borrows a Product-Liability Model for a Technology That Is Not a Product

ATB's Alleged Breach Exposes the Gap Between Ukraine's Wartime Cyber Defence and Its Consumer Data Rules

TikTok's Surrender on the £12.7m Fine Settles Age Assurance for Under-13s, and the Hard Questions Now Move to Teenagers

OpenAI's textGrain Watermark Shows Why the AI Act's Text-Marking Duty Should Be Judged by a Standard of Effort, Not Perfection

Turkey's Article 8/A Blocking Orders Have Become Routine Censorship, Not an Emergency Brake

Malaysia's AI Data Framework Will Matter for Facial Recognition Only If It Reaches the Public Sector

The Dutch DUV Licensing Gap Is Real, but a MATCH Act Ultimatum Is the Wrong Fix

Estonia's Cyber Doctrine Is Right to Bet on Rehearsed Fallbacks, but Machine-Speed Defence Needs Hard Legal Limits

Singapore's Teen Social Media Rules Are Well Targeted, but Mandatory Age Checks Carry the Real Risk