REVIEW 11 cited by
Raising the Cost of Malicious AI-Powered Image Editing
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Raising the Cost of Malicious AI-Powered Image Editing
read the original abstract
We present an approach to mitigating the risks of malicious image editing posed by large diffusion models. The key idea is to immunize images so as to make them resistant to manipulation by these models. This immunization relies on injection of imperceptible adversarial perturbations designed to disrupt the operation of the targeted diffusion models, forcing them to generate unrealistic images. We provide two methods for crafting such perturbations, and then demonstrate their efficacy. Finally, we discuss a policy component necessary to make our approach fully effective and practical -- one that involves the organizations developing diffusion models, rather than individual users, to implement (and support) the immunization process.
Forward citations
Cited by 11 Pith papers
-
I2VShield: An Efficient Proactive Defense Framework against DiT-based Image-to-Video Models
A text-adaptive generator trained to disrupt diffusion-transformer cross-attention shields reference images from video-generation misuse at low online cost.
-
Cross-Branch Conflict as a Shield: Safeguarding Facial Identities in Unified Multimodal Image Editing
CCS jointly perturbs the ViT and VAE pathways and reduces their CKA agreement, causing unified multimodal image editors to lose facial identity.
-
IDDM: Identity-Decoupled Personalized Diffusion Models with a Tunable Privacy-Utility Trade-off
IDDM immunizes authorized personalized diffusion models so public generations remain high-quality while identity linkability to face recognizers is reduced with a tunable privacy-utility knob.
-
Cross-Branch Conflict as a Shield: Safeguarding Facial Identities in Unified Multimodal Image Editing
CCS is an adversarial image protection method that disturbs both the ViT and VAE branches of unified multimodal editing models and disrupts their cross-branch agreement, suppressing identity-preserving edits.
-
Defending from GeoLocalization through Adversarial Road Trips
RoadTrip Attack uses beam search over adaptive geographic intermediate targets to produce stronger, more transferable, lower-visibility adversarial examples against retrieval-based image geolocalizers than PGD, FGSM, ...
-
Closed-Form Concept Erasure via Double Projections
A training-free double-projection linear transformation erases target concepts from generative models by computing a proxy projection then applying a constrained update in the left null space of known directions.
-
SyncBreaker:Stage-Aware Multimodal Adversarial Attacks on Audio-Driven Talking Head Generation
SyncBreaker jointly attacks image and audio streams with Multi-Interval Sampling and Cross-Attention Fooling to degrade speech-driven talking head generation more than single-modality baselines.
-
SyncBreaker:Stage-Aware Multimodal Adversarial Attacks on Audio-Driven Talking Head Generation
A multimodal adversarial attack using stage-sampled image nullification and cross-attention flattening degrades lip-sync and facial dynamics in Hallo-based talking-head generation.
-
DECAF: De-Clustering for Adaptive Representational Unlearning
DECAF is a forget-only unlearning method that adds input noise, suppresses the forget-class probability, and diversifies outputs, achieving 0.10% forget accuracy and 79.4% retain accuracy on CIFAR-10/ResNet-18 while d...
-
VOID: Defeating Unauthorized Mimicry in Latent Diffusion Models
VOID defeats mimicry in LDMs via stochasticity manipulation in the diffusion pipeline, raising average FID from 113 to 365 across evaluations.
-
Bypassing Copyright Protection in Diffusion-based Customization via Two-Stage Latent Feature Optimization
TS-LFO is a two-stage latent feature optimization method that bypasses state-of-the-art copyright defenses in diffusion-based image customization by restoring semantic consistency in latent space.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.