buildConfirmed4 publishers Google opened its SynthID Detector to everyone, checking files for the watermark it has put on more than 180 billion images and videos. Publishing teams get usable evidence from a hit, but the check sees only generators that apply SynthID, so a clean file proves little.
Perspective Coverage
4 publishers
- Builder
- Builder 25%
- Operator
- Operator 53%
- Investor
- Investor 22%
Reality
- Evidence45
- Adoption55
- Hype gap+20
- Incentives70
- Confidence60
buildConfirmed27 publishers OpenAI will watermark ChatGPT and Codex text for EU users, but API customers worldwide get textGrain only if they switch it on from October 5. Teams building on the API now decide for themselves whether to mark output, while Anthropic applies Claude's watermark everywhere.
Perspective Coverage
28 publishers
- Builder
- Builder 34%
- Operator
- Operator 51%
- Investor
- Investor 15%
Reality
- Evidence62
- Adoption
- Not scored
- Hype gap+8
- Incentives72
- Confidence66
buildConfirmed6 publishers Google shipped two text-to-speech models with the same per-line performance control and put voice creation on only one of them, so the tier decision depends on whether a project has to invent a voice at all.
Perspective Coverage
7 publishers
- Builder
- Builder 64%
- Operator
- Operator 23%
- Investor
- Investor 13%
Reality
- Evidence62
- Adoption30
- Hype gap+25
- Incentives58
- Confidence66
Nikon is re-reviewing its Small World in Motion winner after scientists said the microscope video shows structures that do not occur in biology. Its ruling on whether neural-network coloring counts as banned generative AI rests on documents the entrant supplied after he won.
Perspective Coverage
4 publishers
- Builder
- Builder 30%
- Operator
- Operator 64%
- Investor
- Investor 6%
Reality
- Evidence60
- Adoption
- Not scored
- Hype gap+25
- Incentives55
- Confidence60
buildOne report1 publisher GitHub credits broader counting rules for most of its jump to 708 government takedown requests
GitHub received 708 government takedown requests in the first half of 2026 against 98 in all of 2025, a rise it attributes mostly to new counting rules. Two California AI provenance bills from the same season bear on how open source code is licensed and hosted.
Reality
- Evidence45
- Adoption
- Not scored
- Hype gap+10
- Incentives75
- Confidence40
Anthropic says it is watermarking Claude text to satisfy the EU AI Act Transparency Code and plans a detection API. The exposure sits with teams shipping generated work as their own.
Perspective Coverage
7 publishers
- Builder
- Builder 37%
- Operator
- Operator 44%
- Investor
- Investor 19%
Reality
- Evidence48
- Adoption45
- Hype gap+20
- Incentives62
- Confidence55
buildConfirmed4 publishers Anthropic's Article 50 compliance mark is applied during decoding, so text Claude only edited is flagged too. There is no opt-out, no visibility, and no way to audit coverage.
Perspective Coverage
4 publishers
- Builder
- Builder 38%
- Operator
- Operator 55%
- Investor
- Investor 7%
Reality
- Evidence62
- Adoption60
- Hype gap+25
- Incentives55
- Confidence58
The Vine successor from Evan Henshaw-Plath enforces its human-only rule by refusing uploads entirely. The interesting question is what that constraint costs in production and in moderation.
Reality
- Evidence55
- Adoption15
- Hype gap+20
- Incentives60
- Confidence60
buildOne report1 publisher Sparkpix.ai routes detector uploads around its own WebP re-encoder to keep AI labels intact
Sparkpix.ai's detector stores uploads up to 10 MB byte-for-byte because its own WebP re-encoder strips the C2PA and IPTC labels it reads. Chat apps and screenshots lose the same labels, so a missing label clears nothing.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap0
- Incentives75
- Confidence55
buildConfirmed3 publishers A reverse engineer says Cocreator embeds a 16-byte GUID handed out by a Microsoft moderation endpoint into images the NPU generates locally. Local inference and private inference are not the same purchase.
Perspective Coverage
3 publishers
- Builder
- Builder 48%
- Operator
- Operator 40%
- Investor
- Investor 12%
Reality
- Evidence58
- Adoption50
- Hype gap+20
- Incentives
- Not scored
- Confidence60
Gemini 3.8 Flash TTS and Flash-Lite TTS arrive with matching APIs, more than 2,000 stock voices and a path from a 30-second sample to a replica voice. Securing the speaker's permission is the customer's job.
Reality
- Evidence34
- Adoption20
- Hype gap+32
- Incentives72
- Confidence55
buildOne report1 publisher Google introduces two Gemini 3.8 Live audio models, one for high-volume chats and one for complex reasoning
Google positions one live model for high-volume conversation and the other for reasoning that continues while the user is still talking. The product page describes both without performance measurements, token costs or latency figures.
Reality
- Evidence22
- Adoption10
- Hype gap+15
- Incentives70
- Confidence35
The tags Apple made optional in March are now mandatory for every content provider. The company that built AI detection in-house is asking suppliers to certify their own catalogues instead.
Perspective Coverage
3 publishers
- Builder
- Builder 27%
- Operator
- Operator 46%
- Investor
- Investor 27%
Reality
- Evidence72
- Adoption38
- Hype gap+24
- Incentives58
- Confidence66
buildOne report1 publisher Anthropic's text watermark rides out of the model into every Claude API response
The mark is applied during generation rather than at an interface, so anything built on the Claude Platform API returns marked text by default. Reading the mark stays behind a private detection API with a registration list.
Reality
- Evidence34
- Adoption28
- Hype gap+14
- Incentives76
- Confidence36
Amplitude now has to prove it filmed its own trailer
The Humankind 2 trailer drew generative-AI accusations before Amplitude could show anything from the shoot. A location credit is not proof, and a flat denial would carry the same problem: that gap is the standing risk now for anyone publishing video.
Reality
- Evidence38
- Adoption
- Not scored
- Hype gap+32
- Incentives66
- Confidence44
The EU AI Office can now fine and recall. The evidence has to come from somewhere
From 2 August the office can demand documents, run evaluations and push models out of a 450-million-person market. What it cannot yet do is measure frontier models without their makers' help.
Reality
- Evidence42
- Adoption38
- Hype gap+21
- Incentives62
- Confidence41
Spider-Man art book: when clone-stamping reads as AI, a policy is not evidence
Marvel's Brand New Day concept art book was accused of generative AI. The publisher says the culprit was photobashing and the clone tool, and nobody outside the room can check either claim.
Reality
- Evidence22
- Adoption18
- Hype gap+34
- Incentives68
- Confidence30
buildOne report1 publisher A five-hour watermark remover, 14,000 stars, and no detector to test either side
Anthropic says its Claude text watermark can survive editing; Guillaume Meyer's five-hour remover says otherwise. With no public detector, nobody outside Anthropic can settle it.
Reality
- Evidence42
- Adoption38
- Hype gap+22
- Incentives61
- Confidence46
The AI writing policy is a coordination cost showing up on the wrong line item
Clay and Polarstep have both published rules for AI-written internal documents. The interesting part is the arithmetic Synthesia's cofounder used to justify his own memo.
Reality
- Evidence50
- Adoption36
- Hype gap+18
- Incentives60
- Confidence45
buildOne report1 publisher A 14,000-star watermark remover, and no detector to test it against
Anthropic described statistical text watermarking for Claude on August 14 but has published no detector, so neither the mark's durability nor its removal can be independently checked.
Reality
- Evidence44
- Adoption55
- Hype gap+42
- Incentives72
- Confidence49