OpenAI launches "AI Photo Recognition Tool" — Influencers Can No Longer Hide Their Photo Edits

📄Full Article· Automatically extracted by trafilaturaGemini 翻譯1280 words

OpenAI announced the official integration of SynthID invisible watermarking technology developed by Google DeepMind, embedding visually imperceptible identifiers into the pixels of every image generated by ChatGPT, Codex, and the API. These identifiers persist even after secondary processing such as screenshots and resizing. Combined with the existing C2PA Content Credentials metadata, this forms a dual-layer defense mechanism of "metadata + invisible watermark." Launched simultaneously is (Previously: Is OpenAI requiring KYC? You must upload ID documents and a personal headshot to use the gpt-image-1 new image generation model API) (Background context: OpenAI launches ChatGPT Images 2.0: major evolution in text rendering, supports multi-image generation, but "Chinese generation" still stuck) Is that selfie sent by your romantic interest actually them or AI-generated? Now there's an official tool to check. The public verification website launched by OpenAI is straightforward to operate: upload an image, and the system automatically scans whether OpenAI's Content Credentials digital certificate or SynthID invisible watermark is embedded, returning results within seconds. Currently in public preview, it can only identify images generated by OpenAI's own products (ChatGPT, Codex, API). Cross-platform verification is listed in future plans. However, the real technical highlight isn't the website itself, but rather the underlying dual-layer defense mechanism. Public verification tool link here: openai.com/verify Since 2024, OpenAI has been attaching C2PA (Coalition for Content Provenance and Authenticity) metadata to AI-generated images, recording the image's "birth certificate"—who generated it, with what model, and when it was produced. The problem is that metadata gets stripped during processes like screenshotting, forwarding, format conversion, and social platform compression, making it easily ineffective in practice. SynthID, this technology developed by Google DeepMind, can write watermarks directly into the pixel layer of images, completely invisible to the naked eye, but readable by dedicated detectors. According to OpenAI's official blog, SynthID is designed to survive common secondary processing such as screenshots, cropping, and resizing, precisely compensating for C2PA's fragile link in the propagation chain. According to Digital Trends, Kakao, ElevenLabs, and Nvidia have also adopted the SynthID standard simultaneously, and Google has further expanded SynthID verification functionality from the Gemini application to Google Search and the Chrome browser, allowing users to see in search results whether an image is AI-generated or camera-captured. OpenAI also acknowledges that no detection method is foolproof, and currently it can only detect images made by its own models. How to use OpenAI's image verification tool? Go to openai.com/verify and upload an image. The system automatically scans whether it contains C2PA Content Credentials or SynthID invisible watermark. Currently in public preview, only supporting identification of images generated by OpenAI products. What's the difference between SynthID invisible watermark and metadata? C2PA metadata records the complete generation context of an image but may be stripped during screenshotting or forwarding; SynthID writes the watermark into the pixel layer, invisible to the naked eye, capable of surviving secondary processing such as screenshots and resizing. The two complement each other to form a dual-layer defense.

Data Status✓ Full text extractedRead Original (動區 BlockTempo)

🔍Historical Similar Events· Keyword + Asset Matching6 items

2026-05-21

OpenAI Codex officially lands on mobile! ChatGPT App can directly write code and review code

Similarity 120%關鍵字 openai同分類 zh

2026-05-21

OpenAI's reasoning model cracks an 80-year-old math puzzle, endorsed by three top mathematicians

Similarity 120%關鍵字 openai同分類 zh

2026-05-21

Sam Altman drops a bombshell on Silicon Valley startups: swap equity for $2 million in OpenAI Tokens

Similarity 120%關鍵字 openai同分類 zh

2026-05-20