Tech News from Simmons Systems
  • Home
  • About
google photos

Google Photos Adds Hidden AI Edit Labels Without Visible Watermarks

Google Photos is introducing a new "Edited with Google AI" disclosure in the Details tab of modified images starting next week. However, the lack of visible in-frame watermarks means the AI edits remain largely invisible at first glance.

24 Oct 2024 • 1 min read

Starting next week, the Google Photos app displays a new disclosure for images modified by AI tools like Magic Editor, Magic Eraser, and Zoom Enhance. Users find an "Edited with Google AI" label when they scroll to the bottom of the Details section for a specific photo. Google states this addition aims to improve transparency following the rollout of its Pixel 9 smartphones.

Despite this update, AI modifications remain completely invisible at a glance because Google does not place visual watermarks directly on the image frame. If a person encounters an AI-edited picture on social media or in a text message, they see no immediate indication that the image is synthetic. The company also updates metadata for non-generative tools like Best Take and Add Me, but keeps those labels out of the visible Details tab.

The subtle placement of these labels fails to address the main criticism regarding Google's AI editing suite, as most people do not check metadata or details tabs for online images. A Google communications manager tells TechCrunch that this work is not done and the company continues to evaluate additional solutions for greater transparency. Google does not rule out adding visible watermarks in the future as it gathers more user feedback.

Read More at the original source →

Hugging Face Breached by Autonomous AI Agent in Unprecedented Attack

Hugging Face, a major open-source AI platform used by over 50,000 organizations, discloses a significant breach in which attackers leverage an autonomous AI agent to infiltrate its production infrastructure. The intruders exploit two code-execution vulnerabilities within the platform's data-processing pipeline by using a malicious dataset, allowing them
21 Jul 2026 1 min read

Popular AI Coding Tools Vulnerable to Sandbox Escape Attacks

Security researchers at Pillar Security uncover serious sandbox escape vulnerabilities in four major AI coding agents: Cursor, OpenAI's Codex, Google's Gemini CLI, and Antigravity. Rather than attacking the sandboxes directly, the researchers exploit a subtle trust gap. The AI agent stays safely inside its sandbox and
20 Jul 2026 1 min read

OpenAI Pushes for Restrictions on Chinese Open-Weight AI Models

The impressive capabilities of Moonshot's Kimi K3, the largest open-weight large language model from Chinese lab Moonshot, ignite a fierce debate about the future of AI economics and technology. OpenAI's head of strategic futures, Dean W. Ball, argues that the US government should foster regulatory fear
20 Jul 2026 1 min read
Tech News from Simmons Systems © 2026
  • Sign up