AI Aware

Most AI detectors focus on one format. AI Aware covers all of them. The London deep tech startup built a single platform that identifies AI-generated and manipulated content across text, images, video, and audio, with each modality drawing on an ensemble of detection models rather than a single algorithm.

The text detection goes beyond flagging fully AI-written content. It's specifically trained to catch hybrid material: pieces that started as AI output and were then edited by hand, or human writing that was adjusted with AI assistance. Results appear at the sentence and paragraph level with calibrated confidence scores, not a binary pass/fail verdict.

The video and audio detection targets deepfakes and voice cloning, analysing facial inconsistencies, temporal anomalies, and audio synthesis artifacts. Image analysis examines pixel-level signatures left by generative models. The research behind the platform comes from co-founder Dr. Tillman Weyde, a Computer Science researcher at City St George's University London. The company has received Innovate UK funding, awarded to roughly 5% of applicants.

The platform claims to outperform established competitors including Originality, GPTZero, ZeroGPT, and Winston on accuracy, precision, and recall. It reportedly detected DeepSeek with 99.983% accuracy without having been specifically trained on it.

Target users include universities validating student submissions, legal teams authenticating evidence, media organisations fact-checking source material, HR departments reviewing candidate work, and financial services firms flagging document manipulation. Formal partnerships with UKRI and the UK All-Party Parliamentary Group on AI give it institutional credibility that few early-stage detection tools have.

A free trial is available without creating an account. Enterprise API access and tailored integrations are available on request.