Research

The broader goal of my research is to build AI systems that are inherently trustworthy and interpretable. My work generally falls into two areas outlined below.

TBD Motivating Sentence

Specific topics of interest include:

  • AI interpretability;

  • datasets and benchmarks;

  • membership inference and model auditing.

Computer Vision >

Media Forensics

TBD Motivating Sentence

Specific topics of interest include:

  • deepfake detection;

  • media provenance;

  • narrative

AI Safety >

Interpretable & Trustworthy AI

2025 Publications

* Equal contribution. Not necessarily chronological.

Uncovering Conceptual Blindspots in Generative Image Models Using Sparse Autoencoders

Bohacek M.*, Fel T.*, Agrawala M., Lubana E. Under Review.
Project Paper Code Weights — Method for identifying blindspots in T2I models: concepts that the model was trained on but can’t generate.

Human Action CLIPS: Detecting AI-generated Human Motion

Bohacek M. & Farid H. IJCAI-W 2025.
Paper Dataset — This paper proposes a method for distinguishing real and fake (T2V) video using multi-modal semantic embeddings, evaluated on DeepAction, a new dataset of real and AI-generated human motion.

DeepSpeak Dataset v1.0

Barrington S., Bohacek M., and Farid H. ArXiv, abs/2408.05366.
Paper Dataset — This paper introduces DeepSpeak, a large-scale dataset of real and deepfake footage designed to support research on detecting state-of-the-art face-swap and lip-sync deepfakes.

Lost in Translation: Lip-Sync Deepfake Detection from Audio-Video Mismatch

Bohacek M. & Farid H. CVPR-W 2024.
Paper — Method for detecting lip-sync deepfakes by comparing mismatches between audio transcription and automated lip-reading.

Nepotistically Trained Generative-AI Models Collapse

Bohacek M. & Farid H. ICLR-W 2025.
Paper — This paper demonstrates how some generative AI models, when retrained on their own outputs, produce distorted images and struggle to recover even after retraining on real data.

Dataset of News Articles with Provenance Metadata for Media Relevance Assessment

Peterka T. & Bohacek M. ACL-W 2025.
Project Paper Code Data — We introduce a new benchmark for assessing relevance of media in news articles through provenance.

Has an AI Model Been Trained on Your Images?

Bohacek M. & Farid H. Under Review.
Paper Dataset — This paper introduces a method to determine whether a generative AI model was trained on a specific image, helping to address concerns around fair use and copyright in generative AI.

Synthetic Human Action Video Data Generation with Pose Transfer

Knapp V. & Bohacek M. CVPR-W 2025.
Project Paper Code Data — We show that synthetic data of human motion improves performance of action classification and understanding.

Can Pose Transfer Models Generate Realistic Human Motion?

Knapp V. & Bohacek M. FG-W 2025.
Project Paper Code Data — We create a framework for evaluating pose transfer with human participant studies. We demonstrate it on SOTA methods, outlining productive directions for future research.