Resemble AI has released a deepfake threat report alongside two tools aimed at enhancing the verification of digital media in real-time. This announcement comes amid rising concerns over AI-generated content infiltrating social media and other digital platforms. The company highlights estimates from Europol indicating that up to 90% of online content could be AI-generated by the end of 2026.
The newly launched products include a Google Chrome extension that scans images, videos, and audio content, and an X bot that allows users to check the authenticity of suspicious posts without navigating away from the platform. For corporate clients, Resemble AI introduced three features: multimodal watermarking, a zero-retention mode for sensitive media, and a reverse image search tool designed to unearth synthetic content lacking a prior digital footprint.
The deepfake threat report is based on a proprietary database of verified incidents aggregated from global media. Resemble AI reported 1,567 unique deepfake incidents documented in 2025, derived from 3,253 news stories. Each incident was classified by attack type and targeted demographic. Notably, non-consensual intimate imagery and child sexual abuse material made up 20% of the verified cases. The report also revealed that generative AI deepfakes were linked to approximately USD $1.3 billion in confirmed fraud losses, with 80% of incidents not disclosing any financial damages.
Furthermore, the study indicated that the average corporate deepfake incident remains in public discourse for 3.5 years, suggesting that reputational harm can endure long after the event has receded from immediate attention.
The Chrome extension is designed for ease of use, allowing users to scan media on websites with a single click. Results are displayed using a color-coded badge system—green for authentic content, red for AI-generated material, and yellow for uncertain results. The tool also offers frame-by-frame analysis for videos and segment scoring for audio, functioning across various websites and social media platforms, including X, Reddit, Instagram, TikTok, Facebook, LinkedIn, Vimeo, and Twitch.
In tandem, the X bot enables users to assess content within posts on the platform. By tagging the @resemble_detect account with the phrase “is this fake?”, individuals can request an automated scan of images or videos, receiving results directly in the thread. This feature is particularly aimed at journalists, researchers, and the general public looking to evaluate potentially misleading content without exiting the platform.
For enterprises, the multimodal watermarking feature provides a method to sign content at the creation stage across different media formats. By embedding invisible signatures into files, the system aims to establish a chain of custody for digital assets. Additionally, a zero-retention mode has been introduced for businesses in regulated sectors such as finance and healthcare, ensuring that sensitive media is analyzed and immediately purged from the cloud.
The reverse image search tool targets what Resemble AI refers to as “zero-day” synthetic media, which lacks existing digital trails. This tool searches the web for matching images, identifies known debunked content, and traces source material to help pinpoint media that does not conform to established statistical patterns.
In a statement regarding these developments, Zohaib Ahmed, Chief Executive Officer and Co-Founder of Resemble AI, emphasized the broadening scope of challenges presented by synthetic media. “For years, the industry focused on making AI-generated voice, image and video more realistic,” Ahmed stated. “We started by building voice AI models, so we understand how these systems work and how they can be weaponised. Multimodal generative AI security is now foundational for enterprises, employees, and everyday people trying to navigate a world where more content is now synthetic.”
Founded in 2019, Resemble AI has been at the forefront of developing both generative and detection technologies for synthetic media across audio, video, and image formats. The firm reports that its open-source text-to-speech model has surpassed 5 million downloads on Hugging Face, and its detection model has been trained using data from over 160 AI models. As the digital landscape evolves, tools like those developed by Resemble AI may prove vital in maintaining the integrity of information shared across platforms.
See also
Sam Altman Praises ChatGPT for Improved Em Dash Handling
AI Country Song Fails to Top Billboard Chart Amid Viral Buzz
GPT-5.1 and Claude 4.5 Sonnet Personality Showdown: A Comprehensive Test
Rethink Your Presentations with OnlyOffice: A Free PowerPoint Alternative
OpenAI Enhances ChatGPT with Em-Dash Personalization Feature
