The European Commission is continuing its investigation into Elon Musk’s X over concerns its AI tool Grok was used to create sexualised images of real people. This follows a similar announcement in January from the UK watchdog Ofcom.
If Musk’s site is found to have breached the rules of EU’s under the Digital Services Act, the Commission could fine the company up to 6% of its global annual turnover.
Meanwhile, other technology firms are adopting a more progressive stance to deepfake. OmniSpeech, the Maryland-based tech startup who earlier this year introduced the AI Detect™ plugin to the Zoom Marketplace, is set to bring their technology to Google Chrome.
Deepfake detection is the process of determining whether a video, audio clip, or image has been artificially manipulated or created using AI.
With AI voice clones getting more realistic by the day and deepfake content proliferating rapidly, there is a clear and present need for reliable tools that help guard against this strange new security threat. Google is one example of a technology firm seeking to alert users to the rising tide of deepfake imagery and voices.
A rapidly evolving landscape
AI Detect™ for Google Chrome employs a proprietary Gen 2 model that is optimised for previously unseen data sets. This approach ensures a high degree of analytical accuracy irrespective of generator, language, codec, or environment, setting OmniSpeech apart from other companies who tend to rely on brittle signature-based approaches which can’t keep up with the rapid pace of deepfake technology evolution.
The plugin can analyse voice audio in real-time from any source including live video calls and pre-recorded clips, making it a highly flexible tool for any number of applications. OmniSpeech also offers API integration for seamless interfacing with cybersecurity and fraud-prevention platforms, enterprise communication tools, and any other application requiring voice authenticity assurance.
The technology is aimed at:
- Journalists & researchers: To verify the authenticity of online interviews, leaked recordings, and user-generated audio before publication.
- Business professionals & executives: To add an additional layer of trust during browser-based calls, negotiations, and identity-sensitive conversations.
- Compliance & security teams: To detect voice spoofing and impersonation attempts across cloud tools, web platforms, and for legal evidence.
- Educators & students: To evaluate lectures, submissions, or online media for AI-generated speech.
- Consumers: To scan suspicious voice messages, scam videos, or viral audio encountered on social media, YouTube, or websites directly in Chrome.
The deepfake detection tool functions by analysing visual, audio, and metadata signals to reveal whether a piece of media has been artificially generated. The software does so by looking for irregular patterns that do not appear in genuine human speech, movement, or image formation.
