Вход на сайт

Просмотр новости

Найдите то, что Вас интересует

Google makes SynthID AI detector publicly available

Дата публикации: 08-10-2026 07:47:20

Google has, to date, watermarked more than 180bn images and videos, it said.
Read more: Google makes SynthID AI detector publicly available


Основное содержимое страницы с новостью.

Google has, to date, watermarked more than 180bn images and videos, it said.

Three years after first launching the tool, Google is now making its generative AI detector available to the public globally, as AI-created content becomes increasingly difficult for humans to discern.

Google has, to date, watermarked more than 180bn images and videos, along with 240,000 years’ worth of audio content, it said. That number stood at around 10bn pieces of content in May 2025.

Users can now use the SynthID detector to identify certain types of AI-generated content created by Google or its affiliated partners, including OpenAI and Nvidia. They will be similarly able to check AI content from Apple’s tools soon, the company said.

Users will be able to utilise the tool alongside Google’s other built-in verification features in Search, the Gemini app, and Chrome, it said.

AI watermarking technology is not entirely reliable, researchers, AI companies and lawmakers all agree.

The SynthID system, for example, can only detect if content was generated using SynthID-enabled systems, but does not reveal other vital information, such as creator identity, tool name, edit history or timestamps, leading to challenges discerning human input into content generated using AI.

SynthID was first launched in 2023, and two years later, media professionals were given access to an early version to help detect AI content.

Earlier this week, OpenAI rolled out text watermarking in the EU; in August, Anthropic pledged the same within the bloc from its newer models.

“Text watermarking and detection remain early technologies with significant limitations,” OpenAI said earlier this week.

Shorter passages and AI-generated output edited by humans are more difficult to detect for watermarking, the company clarified. According to its testing, detection rates plummet from close to 100pc when generated texts are unedited to less than 20pc when one-quarter of the words are changed.

Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.

Схожие новости

#Наименование новостиТональностьИнформативностьДата публикации
1Google Opens SynthID Detector to the World as AI Watermarking Gains Industry Allies09.2107-10-2026
2Google Releases SynthID: Free Online Tool Detects AI-Generated Images and Videos011.6408-10-2026
3Google's SynthID Detector can now tell you if an image, video, or audio is AI-generated014.0607-10-2026
4Google DeepMind Adds Watermarks to AI-Designed Proteins Without Breaking Them011.3801-10-2026
5Google abre a todo el mundo su herramienta para detectar si una imagen, vídeo o audio ha sido creado con IA08.407-10-2026
6Googles KI-Detektor SynthID ist jetzt für (fast) alle zugänglich013.8207-10-2026
7Google pone gratis la herramienta definitiva para saber al instante si una foto o un vídeo de internet está hecho con IA05.4807-10-2026
8Google придумала «ватермарки» для белков, созданных ИИ08.4602-10-2026
9OpenAI to watermark AI-generated text in EU014.4906-10-2026

Классификация: Пресс-релизы. Схожих патентов: 0. Схожих новостей: 9. Тональность: 0. Информативность: 9.11. Источник: www.siliconrepublic.com.