SynthID has been named one of @TIME's Best Inventions of 2026. With over 180 billion images and videos already watermarked, it’s making it easier for people to distinguish AI-generated content from other media → goo.gle/4jsZxWF

Oct 15, 2026 · 12:53 PM UTC

186
62
56
638
74,630
RelevantRecentLikes
Evan Kirstel #B2B #TechFluencer
@EvanKirstelOct 16
Congrats on the TIME nod for SynthID. Watermarking AI content at scale is one of those quiet pieces of infrastructure every enterprise will end up needing as synthetic media becomes the norm.
1
133
180B watermarked and my feed is still 90% slop
1
167
180 billion watermarked is a wild scale
1
190
every time my elderly parents send me a video of a cute animal, it's so obviously ai. maybe i can get them to use this before they forward anything.
1
195
synthid on 180 billion images is huge can anyone check if an image has it or only google ?
1
3
395
A missing watermark isn't proof of anything
1
3
492
180 billion watermarks and my feed is still all ai slop
1
3
246
180 billion watermarked and my feed still feels mostly fake
2
138
180 billion is a lot of pixels to watermark. how does it hold up after a screenshot or a recompress
3
344
Congrats and fuck TIME if this is the best invention they could think of
1
4
441
日本語での詳細解説はこちら Check out my Japanese breakdown: xitter.cf/airiaiai8/status/21109…
AI生成コンテンツの見分け方がついに実用レベルに。GoogleDeepMindの電子透かし技術「SynthID」がTIME誌のBest Inventions 2026に選出された。 既に1800億枚以上の画像・動画に透かしを埋め込み済みというスケール感がヤバい。 以下、解説 ===↓↓↓ 【SynthIDとは】 Google DeepMindが開発した電子透かし技術。AI生成画像・動画・音声・テキストに、人間の目には見えない形でデジタル署名を埋め込む。生成後に画像を加工・圧縮しても検出可能な耐久性が特徴。 【規模感】 - 既に180億枚以上の画像・動画に適用済み - Google製品(Imagen, Veo等)の生成コンテンツに標準搭載 - ディープフェイクやAI生成フェイク画像の判別に活用され始めている 【なぜ今、評価されたのか】 生成AIの急速な普及で「本物とAI生成物の区別がつかない」問題が深刻化。SynthIDは情報の信頼性を担保するインフラとして機能し始めており、TIME誌が選ぶ「2026年の発明」に選出されたのはその社会的インパクトの証左。 【今後の展開】 GoogleはSynthIDの技術をオープンソース化し、他社のAIモデルでも利用できるよう展開を進めている。業界全体での「AI生成物の透明性確保」の動きが加速しそうだ。
1
79
my art is watermarked by being bad
104
Batch codes for pixels. Food packs have carried them for decades so you can trace where something came from. About time photos showed their paperwork too.
1
261
The Lunar Misfit
@ExiledonMoonOct 15
Google won an award for telling us which pictures Google made up.
1
1
133
useful step, though bad actors will keep trying to evade markers
32
180 billion watermarked and my feed still full of slop lol
40
180 billion watermarked is a huge number. I seal my own videos before release too, so I love seeing provenance become normal.
1
220
180 billion watermarked is the easy number. I'd want the detection rate after a crop, resize, and re-encode, since that's where most images actually end up.
1
180
180 billion and my group chat still forwards the fake ones like breaking news
52
Roughly 1% (or less) of all photos/images estimated to exist worldwide
185
this is a game changer broo
1
A real photo edited with Google's image tools can carry SynthID. So detection can mean an edit, not an invented scene. Google also notes that repeated alterations can defeat detection. No watermark found is not the same as a genuine photo.
2
35
A useful UX test: can a first-time user distinguish 'no SynthID watermark found' from 'this is authentic'? I'd show supported generators and an explicit 'origin unknown' state beside the result. Coverage is part of the answer.
1
46
Does the SynthID check in Gemini show which parts of an image are watermarked, like the Detector does? Because someone could think a real photo is fake just because one small detail was edited with AI.
1
35
Google, Anthropic and OpenAI all watermark chatbot text now, with different reach. Gemini does it in its app and on the web. Anthropic marks every current Claude model by default, API included. OpenAI is rolling it out to ChatGPT and Codex in the EU only, and the API is opt-in.
1
61
180 billion images/videos already watermarked. detector covers openai/nvidia/kakao and soon apple too the "soon apple" line is the interesting part
1
56
180 billion watermarked images and videos is real scale for provenance. The next test is interoperability: watermarks build trust when they can be checked across platforms and survive edits. Congratulations on the TIME recognition.
18
Gemini・Google検索・Chromeに画像や動画、音声をアップロードして「Google AIで作られたか」と聞くと、SynthIDの透かしが入っているか確認できます。SynthID Detectorも誰でも使えるようになっていて、OpenAIやNVIDIA、Kakaoのモデルで作られたものも判定でき、Appleも今後対応予定みたいです。
51
180B items watermarked is the reach. Robustness is the other half: on text, a July 2026 arXiv study saw paraphrasing strip SynthID's mark in 58 of 59 detected cases. Any public numbers for images after crop/re-encode?
2
这里最想确认的是水印的真实存活率:这 1800 亿次标记里,经过压缩、裁剪、转码之后仍能被检出的比例是多少?你们的公开数据是否包含按"攻击类型"拆分的检出率?这会直接改变结论的强度——实验室环境和真实传播链完全是两回事。
8
1800億枚って桁がもうやばい…
7
180B watermarked is a big number. The real test is the other side: how many platforms actually check for it at upload, and how well it survives crops, re-encodes and screenshots?
10
Congrats to the team! Helping people recognize AI-generated content deserves attention alongside the tools that create it.
12
watermarking at that scale is good for creators too. people should be able to tell what's made with AI and still love it
11
watermarking generative media at scale is a superficial band aid that fails to address the core vulnerability of easily manipulated model weights
12