Loading story…
Researcher Bill Swearingen publicly unveiled noRecognition, a project that uses reinforcement learning to generate adversarial visual patterns designed to stop automated surveillance systems from recognizing people, objects, and vehicles while still allowing ordinary video recording. Swearingen said the system ran roughly 31 million tests over the past year and produced increasingly effective patterns by learning from prior detection failures.
In lab testing, Swearingen said the patterns defeated 11 open-source detection algorithms, including software associated with Flock license plate readers, Axon body-worn camera workflows, and Clearview AI-enabled camera pipelines. He also demonstrated the concept at Def Con by covering a 2009 Toyota Yaris with one of the patterns and testing it against a Flock camera, while keeping the most effective designs offline to slow vendor countermeasures and pursuing crowdfunding for derivative products such as shirts, hoodies, and possible vehicle wraps.
Mallory correlates global threat intelligence with your attack surface — know if you’re exposed before adversaries strike.
5 events from the most recent confirmed update back to the earliest known activity.
At Def Con 2026 in Las Vegas, Swearingen publicly presented noRecognition and demonstrated the concept in real-world conditions. He covered a 2009 Toyota Yaris with one of the generated patterns and tested it against a Flock camera, which he said proved effective.
Swearingen said the project was using crowdfunding to support derivative products based on the patterns. Planned items included T-shirts and hoodies, with possible vehicle skins or wraps later.
Swearingen said his lab testing found patterns that defeated all 11 open-source detection algorithms he evaluated. The tested software included algorithms associated with Flock license plate readers, Axon body-worn cameras, and Clearview AI-enabled camera workflows.
Over the course of a year, Swearingen said the system ran roughly 31 million tests to iteratively improve its adversarial patterns using reinforcement learning. He said the model eventually generated new patterns every minute and used failures as feedback for better results.
Bill Swearingen created noRecognition, a reinforcement-learning project that generates visual patterns intended to disrupt automated surveillance detection of people, objects, and vehicles without preventing video recording. He said he began the work out of concern about pervasive surveillance and its effects on privacy and protest participation.
Vulnerabilities, threat actors, malware, products, organizations, and breaches Mallory has linked to this story.
3 references tracked. Mallory keeps watching after this page renders.
scworld.com
Open sourcecyberveille.ch
Open sourcetechcrunch.com
Open sourceMap indicators from this story to your assets and identify affected systems in minutes.
Every observed campaign, victim, and pivot linked to actors named in this story.
Malware, exploits, and IOCs connected to the activity described here.
YARA, Sigma, and Snort rules deployed to your SIEM as soon as they’re published.
Get matching new stories delivered to your team as they break — not the next morning.
Ask questions about this story and take action on the answers.
Other stories Mallory is tracking that overlap on entities, sources, or topic labels.
A range of legal and technical measures are being taken to counter the use and abuse of surveillance and data collection technologies by both government agencies and private corporations. In the United States, privacy advocates and hackers are developing counter-surveillance tools to disrupt automated license plate readers (ALPRs) operated by companies like Flock, which supply footage to law enforcement agencies including ICE. These efforts include adversarial license plate overlays that confuse AI recognition systems and open source mapping of ALPR networks. Additionally, a significant security lapse was discovered in Flock's infrastructure, exposing live camera feeds and logs due to misconfigured, publicly accessible admin interfaces. Legal action is also underway, with organizations like the EFF and ACLU suing municipalities over alleged abuses of surveillance technology. In Texas, a court has issued a temporary restraining order against Samsung, prohibiting the company from collecting smart TV viewing data from consumers without proper consent. The court found Samsung's Automated Content Recognition (ACR) practices deceptive, highlighting concerns about undisclosed data collection and potential access by foreign entities such as the Chinese government. The ruling underscores growing legal scrutiny of pervasive data collection by technology vendors and the opaque enrollment processes that mislead consumers about the extent and use of their personal information. These developments reflect increasing resistance—both technical and legal—to the expansion of surveillance and data harvesting technologies in the US.
Jun 29, 2026
The New York Metropolitan Transportation Authority (MTA) is testing new subway gates that use **AI-powered cameras** to capture short recordings when riders are suspected of fare evasion and to generate a physical description that is transmitted to the MTA, prompting criticism from privacy advocates concerned about persistent monitoring in public transit. The MTA has also solicited vendor input for systems using computer vision and AI to detect “unusual or unsafe behaviors,” reflecting broader growth in surveillance deployments across New York City. In parallel, consumer **AI smart glasses** are re-emerging with built-in cameras and microphones, intensifying concerns that everyday wearables can enable covert recording and downstream biometric identification. Reporting highlighted that footage from *Ray-Ban Meta* smart glasses can be paired with external facial-recognition services to identify strangers, and noted policy issues such as cloud storage of wake-word voice recordings (potentially retained up to a year) and uncertainty about future features like on-device facial recognition; retailers in New York (e.g., Wegmans and others) are also expanding facial-recognition use, underscoring the convergence of AI, biometrics, and surveillance in both public and commercial spaces.
Jun 29, 2026
Researchers at the University of Pisa have developed a novel image signature system that maintains its integrity even after an image is cropped, addressing a major vulnerability in current image authentication methods. This technology allows the original signature to remain verifiable on cropped images by dividing the image into blocks, ensuring that only legitimate edits like cropping are permitted while any manipulation within the blocks invalidates the signature. The innovation aims to help newsrooms and publishers maintain trust in visual content, even after routine editing, and to prevent deepfakes from exploiting weaknesses in image verification. Meanwhile, the UK's Information Commissioner's Office (ICO) has criticized the Home Office for failing to disclose significant biases in police facial recognition algorithms used within the Police National Database. Recent tests revealed that the currently deployed Cognitec FaceVACS-DBScan ID v5.5 algorithm exhibits notable weaknesses in identifying certain demographics under strict verification settings, raising concerns about fairness and transparency. The ICO has demanded urgent clarification from the Home Office, emphasizing the importance of public trust and the need for accountability in the deployment of facial recognition technologies.
Jun 29, 2026