Global Compliance & Risk Management

Field notes on the operational layer of platform governance.

Visual guides to the regulation. Reports on the implementation. Writing from inside seven years of trust & safety, online safety, and compliance work.

Browse the library What's new Subscribe
13 frameworks · 7 dispatches · 5 jurisdictions · Last updated 21 May 2026
§ Most recent

Just added.

§ The Library · 13 frameworks

Visual guides to the laws, bills and regulations reshaping platforms.

Each guide breaks the statute down to its operational implications. Color-coded by jurisdiction.

Filter
AI · cross-jurisdictional 2
EU 4
UK 1
US 4
APAC 2
§ Dispatches · 7 pieces

Reports, field notes, and live trackers.

Longer-form analyses and ongoing monitoring of where the laws above are being applied, contested, or quietly worked around. Filed from inside the operational layer.

Field note · May 20264 min read

The Mood Among New Grads

A short field note from a research conversation with fresh Cornell graduates on how AI is landing with the Class of 2026. The word that kept coming up was "robbed." Filed in the same week as Meta's 8,000-person layoff and a string of commencement-speech boos at Arizona, UCF, Middle Tennessee State, and Glendale Community College.

Read the field note
Field note · May 202618 min read

The Deployer's Frame

The UK CMA's 9 March 2026 framework for businesses deploying AI agents in consumer-facing roles — the first guidance of its kind from any major consumer-protection authority. The four operational requirements, the four worked examples, the legal stack underneath (CRA 2015, CPRs 2008, CCAR Regs 2013, DMCCA 2024), and why the deployer carries everything — including what was supplied by a third party.

Read the field note
Field note · May 202622 min read

The Two-Consent Problem

Agentic AI and the gap between user authorization and platform authorization. After Amazon v. Perplexity (9 March 2026) and the CMA's same-day guidance, a three-layer authorization stack, three peer-reviewed sources, and a fifty-state US patchwork pulling against federal preemption.

Read the field note
Report · May 202624 min read

The AI Reckoning

Safety, security, adoption, and the real cost of moving fast. Data, timelines, case studies, and research from inside the industry on what the wave of frontier-AI deployment has actually produced so far.

Read the report
Field note · May 202619 min read

The Camera Roll Cloud

Meta's opt-in camera roll suggestions launched in the EU and UK on 16 April 2026. The two-toggle anatomy, the AI Terms underneath, regional carve-outs (Illinois & Texas excluded), peer comparison against Apple, Google and Snapchat, and the 30,000-photo insider breach disclosed nine days before launch.

Read the field note
Live tracker · Quarterly6 min read

What Ofcom is saying about the OSA

Each quarter, Ofcom publishes an industry bulletin signalling which Parts of the Online Safety Act it's animating through codes, guidance, supervision, or enforcement. 5 editions tracked, plotted against the Act coming alive.

Open the tracker
Compliance report · Mar 20267 min read

OnlyFans · OSA Compliance Report

Applicability, Part 5 vs Part 3 duties, age assurance, CSEA reporting, Category 1 obligations. The practical walk-through of one of the UK Online Safety Act's hardest test cases.

Read the report

What I'm reading this week.

Week of 18 May 2026
The EU simplified its toughest AI law: what changed and why it matters
High-risk AI obligations under Annex III pushed from summer 2026 to December 2027. Embedded AI in medical devices and industrial machinery pushed to August 2028. The most consequential AI Act change since the regulation passed.
Euronews · 21 May
Commission seeks feedback on draft guidelines for classifying high-risk AI systems
The Commission opened public consultation on 19 May on how Annex III high-risk systems should be classified. Worth reading the draft itself if you're building anything employment-, education-, or health-adjacent.
European Commission · 19 May
Meta slashes 8,000 jobs as it pivots towards AI
The framing matters: layoffs explicitly funded the AI push in the memo, with another 7,000 staff being redirected into new AI teams. The most explicit "replace, not augment" signal from a major platform so far.
NPR · 20 May
EU AI Act Update: Timeline Relief, Targeted Simplification, and New Prohibitions
Covington's full breakdown of the Omnibus deal. Buried in the simplification narrative: a new prohibition on "nudifier" apps, grandfathering rules for existing GenAI systems, and centralisation of GPAI enforcement in the AI Office.
Covington · 18 May
Implementation of the Online Safety Act
Updated Commons Library briefing on where OSA implementation actually stands. Useful as a single-source check against the press coverage, which has been swinging between "Ofcom is too slow" and "Ofcom is overreaching" depending on the week.
UK Parliament · 19 May
§ Cross-jurisdictional · Key dates

When the rules change.

The dates that matter across the laws this site covers. Past, present, and what's queued next. Filter by jurisdiction or scroll the lot.

§ Next 24 months · Upcoming dates May 2026 → May 2028
Aug 2026 · EU AI Act · full application
Aug 2027 · EU AI Act · Article 6 high-risk
May '26
Nov '26
May '27
Nov '27
May '28
Filter
Aug 2027
EU AI Act · Article 6
High-risk AI systems obligations apply in full
Aug 2026
EU AI Act
Full application of the AI Act outside high-risk Annex III
Aug 2025
EU AI Act · Chapter V
GPAI obligations apply: transparency, copyright, systemic risk
Jul 2025
UK Online Safety Act · ss.11-12
Children's safety duties commence; mandatory age assurance for adult content
Mar 2025
UK Online Safety Act · ss.9-10
Illegal content duties commence; Ofcom Codes of Practice take effect
Feb 2025
EU AI Act · Article 5
Prohibited AI practices ban applies; AI literacy obligations start
Aug 2024
EU AI Act
Regulation (EU) 2024/1689 enters into force
Mar 2024
EU Digital Markets Act
DMA full application for designated gatekeepers
Feb 2024
EU Digital Services Act
DSA full application for all in-scope intermediaries
Oct 2023
UK Online Safety Act
Royal Assent; phased commencement begins under Ofcom
§ Coverage · 5 jurisdictions · 13 frameworks

The structure.

Platform governance at the centre, jurisdictions radiating out, frameworks at the edges. Click any node to open its guide.

§ Live research · Open

The Trust Gap.

A short, anonymous survey on public perceptions of data collection by social media and tech companies. About 10 minutes.

Take the survey
§ About

I'm Lubos.

Seven years working on online safety, trust & safety, and compliance. TikTok, Lockwood Publishing, now Klaviyo. The work that shapes the experience of being online sits below the regulation that frames it, in the policy decisions, the risk thresholds, the escalation chains, and the moderation queues. This site is where I write some of it down.

Read the full bio
Lubos Dusek