An international team of researchers is calling for urgent improvements in deepfake detection technologies after uncovering critical flaws in widely used detection tools.
A study by CSIRO, Australia's national science agency, and South Korea's Sungkyunkwan University (SKKU), assessed 16 leading detectors and found none could reliably identify real-world deepfakes.
Deepfakes are artificial intelligence (AI) generated synthetic media that can manipulate images, videos, or audio to create hyper-realistic but false content, raising concerns about misinformation, fraud, and privacy violations.
CSIRO cybersecurity expert, Dr Sharif Abuadbba, said the availability of generative AI has fuelled the rapid rise in deepfakes, which are cheaper and easier to create than ever before.
"Deepfakes are increasingly deceptive and capable of spreading misinformation, so there is an urgent need for more adaptable and resilient solutions to detect them," Dr Abuadbba said.
"As deepfakes grow more convincing, detection must focus on meaning and context rather than appearance alone.
"By breaking down detection methods into their fundamental components and subjecting them to rigorous testing with real-world deepfakes, we're enabling the development of tools better equipped to counter a range of scenarios."
The researchers developed a five-step framework that evaluates detection tools based on deepfake type, detection method, data preparation, model training, and validation.
It identifies 18 factors affecting accuracy, ranging from how data is processed to how models are trained and tested.
SKKU Professor Simon S. Woo, said the collaboration between CSIRO and SKKU has advanced the field's understanding of detection model vulnerabilities.
"This study has deepened our understanding of how deepfake detectors perform in real-world conditions, exposing major vulnerabilities and paving the way for more resilient solutions," he said.
The study also found many current detectors struggle when faced with deepfakes that fall outside their training data.
For example, the ICT (Identity Consistent Transformer) detector, which was trained on celebrity faces, was significantly less effective at detecting deepfakes featuring non-celebrities.
CSIRO cybersecurity expert, Dr Kristen Moore, said using multiple detectors and diverse data sources strengthens deepfake detection.
"We're developing detection models that integrate audio, text, images, and metadata for more reliable results," Dr Moore said.
"Proactive strategies, such as fingerprinting techniques that track deepfake origins, enhance detection and mitigation efforts.
"To keep pace with evolving deepfakes, detection models should also look to incorporate diverse datasets, synthetic data, and contextual analysis, moving beyond just images or audio."
The paper, SoK: Systematization and Benchmarking of Deepfake Detectors in a Unified Framework, was published in arXiv preprint . The paper has been accepted at IEEE European Symposium on Security and Privacy 2025. This paper is authored by Le, Binh M., Jiwon Kim, Simon S. Woo, Kristen Moore, Alsharif Abuadbba, Shahroz Tariq (CSIRO Technical Lead).
The types of real-world deepfake types that outsmarted detectors
Synthesis:
Synthesis deepfakes generate entirely new synthetic faces using AI-powered Generative Adversarial Networks (GANs) or Diffusion models. They create an artificial identity by blending or generating facial features. These deepfakes are often used for creating realistic looking but non-existent individuals. For instance, a Diffusion model can blend the faces of two famous actors.
Faceswap:
Faceswap deepfakes replace one person's face with another in a video while keeping the original body and background. For example, they can make it look like a celebrity's face is on someone else's body in a video.
Reenactment:
Reeanactment deepfakes transfer facial expressions and movements of one person to onto another's face in a video. This preserves the target person's facial features but alters their expressions. They can be used to convincingly fabricate speeches or reactions that never occurred.
