Provenance Record

Well-structured News piece following the standard Overview / What We Know / What We Don't Know / Analysis format. Clear, neutral prose; appropriate length (676 words) for the News category (400-1200 word range). The Analysis section's framing — that this case targets Meta's core engagement mechanics rather than a discrete compliance gap, unlike the earlier X and Temu DSA fines — is a reasonable, source-supported editorial synthesis rather than an unsupported claim.

All 5 source snapshots read in full from sources/2026-07/eu-commission-finds-instagram-and-facebooks-addictive-design-breaches-digital-services-act/ (source-0.html.gz through source-4.html.gz, all HTTP 200, sha256 hashes verified against manifest.json by decompressing and rehashing each file). source-0 (EC press release, 'Commission preliminarily finds...') confirms the headline claim, the 'infinite scroll, autoplay, push notifications, and highly personalised recommender systems' quote, the 'Meta did not adequately assess the risks...' language, and the Virkkunen quote verbatim. source-1 (EC press release, 16 May 2024) confirms the formal-proceedings opening date and both direct quotes about 'rabbit-hole effects' and 'age-assurance and verification methods' verbatim. source-2 (The Register) confirms the 6% turnover fine cap, the ~$12B estimate, and corroborates the Commission's design-change demands (disabling autoplay/infinite scroll by default, less engagement-focused recommendations). source-3 (Euronews) confirms verbatim the 'disregarded data on how much time teenagers spend... at night', the reels/stories optimisation language, the time-management-tools and parental-controls findings, the $12 billion/€11bn calculation against $201B 2025 revenue, and the €120M X / €200M Temu prior DSA fines. source-4 (Al Jazeera) confirms verbatim the 'autopilot mode' quote, the Ben Walters/Meta spokesperson pushback quote, the two-year-investigation characterization, and the Fabrizio Esposito 'strike at the heart' quote with correct title (NOVA School of Law, Lisbon). No hallucinated quotes, no misattributed specifics, no orphan URLs (all 5 sources array URLs appear in body_markdown and vice versa, confirmed by the automated 'body_sources_match' check).

Every direct quote in the article appears verbatim in its cited snapshot. Every specific figure (6% turnover cap, ~$12B/€11bn estimate, $201B 2025 revenue, €120M X fine, €200M Temu fine, 16 May 2024 proceedings-opening date) traces exactly to its cited source. One minor style note: several passages presented in quote marks (e.g. 'disregarded data on how much time teenagers spend...') reproduce the cited outlet's own reporting prose verbatim rather than a quoted spokesperson statement — this is accurate word-for-word attribution to the outlet, not a hallucination, and is standard practice for lifting exact phrasing from a press release or article. The headline omits 'preliminarily' (as do all four human news outlets' own headlines in the source set), but the summary and 'What We Don't Know' section immediately and repeatedly clarify the finding is preliminary and subject to Meta's response, so this is not materially misleading.

High-quality, thoroughly sourced submission. All five sources personally verified against every claim, quote, and figure attributed to them. No corrections needed. Approved as-is.