Deepfakes and AI-Generated Content — What Detection Actually Tells You in 2026 — Quiz

1.

What is a Generative Adversarial Network (GAN)?

2.

The 'liar's dividend', as described by Chesney and Citron (2019), refers to:

3.

What is the key asymmetry in interpreting a deepfake detection tool's output?

4.

Why do deepfake detectors trained on GAN-generated content perform poorly on diffusion-model content?

5.

According to the ENISA Threat Landscape 2025, which form of synthetic media is highlighted as a growing vector in fraud?

6.

What is the EU AI Act's Article 50 requirement regarding AI-generated content?

7.

Explain what the 'liar's dividend' is and why it represents a risk that is independent of whether any specific deepfake is successfully distributed.

8.

Describe two limitations of current deepfake detection tools that mean they cannot serve as definitive arbiters of whether content is authentic.