Table 8.

A summary of attacks and defenses for diffusion models

Attack/DefenseMethodYearCategorySubcategoryTarget modelsDataset
Adversarial AttackECB (Struppek et al., 2023a)2024Black-boxCharacter-levelStable Diffusion, DALL-E 2, AltDiffusion-m18LAION-Aesthetics v2, MS COCO, ImageNet-V2, self-constructed
CharGrad (Kou et al., 2023)2023Black-boxCharacter-levelStable DiffusionMS COCO, Flickr30k
ER (Gao et al., 2023)2023Black-boxCharacter-levelStable Diffusion, DALL·E 2LAION-COCO, DiffusionDB, SBU Corpus, self-constructed
DHV (Daras and Dimakis, 2022)2022Black-boxWord-levelDALLE-2
AA (Millière, 2022)2022Black-boxWord-levelDALL-E 2, DALL-E mini
BBA (Maus et al., 2023)2023Black-boxSentence-levelStable DiffusionImageNet
RIATIG (Liu et al., 2023b)2023Black-boxSentence-levelDALL·E, DALL·E 2, ImagenMS COCO
QFA (Zhuang et al., 2023)2023Grey-boxSimilarity-drivenStable Diffusionself-constructed
RVTA (Zhang et al., 2024b)2024Grey-boxSimilarity-drivenStable DiffusionImageNet, self-constructed
MMP-Attack (Yang et al., 2024a)2025Grey-boxSimilarity-drivenStable Diffusion, DALL-E 3, Imagine ArtMS COCO
DORMANT (Zhou et al., 2024b)2025Grey-boxDistance-drivenAnimate Anyone, MagicAnimate, MagicPose, MusePose, Champ, MuseV, UniAnimate, and ControlNeXtTikTok, Champ, UBC Fashion, and TED Talks
ATM (Du et al., 2024a)2023White-boxClassifier-drivenStable DiffusionImageNet, self-constructed
FOOLSDEDIT (Zhou et al., 2024c)2024White-boxClassifier-drivenSDEditCelebAMask-HQ, FFHQ
SAGE (Liu et al., 2024h)2023White-boxClassifier-drivenGLIDE, Stable Diffusion, DeepFloydImageNet
SneakyPrompt (Yang et al., 2024j)2023Black-boxTarget External DefensesStable Diffusion, DALL·E 2NSFW-200, Dog/Cat-100
UD (Qu et al., 2023)2023Black-boxTarget External DefensesStable Diffusion, LD, DALL·E 2, DALL·E miniMS COCO
Jailbreak AttackAtlas (Dong et al., 2024)2024Black-boxTarget External DefensesStable Diffusion, DALL·E 3NSFW-200, Dog/Cat-100
Groot (Liu et al., 2024q)2024Black-boxTarget External DefensesStable Diffusion, Midjounery, DALL·E 3self-constructed
DACA (Deng and Chen, 2023)2024Black-boxTarget External DefensesMidjounery, DALL·E 3VBCDE-100, Copyright-20
SurrogatePrompt (Ba et al., 2024)2024Black-boxTarget External DefensesMidjourney, DALL·E 2, DreamStudioself-constructed
PGJ (Huang et al., 2025d)2025Black-boxTarget External DefensesStable Diffusion, DALL-E 2, DALL-E 3, Cogview3, Tongyiwanxiang, Hunyuanself-constructed
R2A (Zhang et al., 2025b)2025Black-boxTarget External DefenseStable Diffusion, FLUX, DALL·E 3, Midjourneyself-constructed
Jailbreak AttackJPA (Ma et al., 2024a)2024Grey-boxTarget Internal DefensesStable Diffusion, Midjourney, DALL·E 2, PIXART-αI2P
RT-Attack (Gao et al., 2024e)2024Grey-boxTarget Internal DefensesStable Diffusion, DALL·E 3, SafeGenI2P, self-constructed
RTSDSF (Rando et al., 2022)2022White boxTarget External DefensesStable Diffusionself-constructed
MMA (Yang et al., 2024i)2024White boxTarget External DefensesStable Diffusion, Midjounery, Leonardo.AiLAION-COCO, UnsafeDiff
P4D (Chin et al., 2024)2024White boxTarget Internal DefensesStable DiffusionI2P, ESD Dataset
UnlearnDiffAtk (Zhang et al., 2023g)2024White boxTarget Internal DefensesStable DiffusionI2P Dataset, ImageNet, WikiArt
ESD (Gandikota et al., 2023)2023Concept ErasureFine-tuningStable DiffusionMS COCO, I2P
SPM (Lyu et al., 2024)2024Concept ErasureFine-tuningStable DiffusionMS COCO, I2P
SDD (Kim et al., 2023)2023Concept ErasureFine-tuningStable DiffusionMS COCO, I2P
AC (Kumari et al., 2023)2023Concept ErasureFine-tuningStable DiffusionMS COCO
ABO (Hong et al., 2024a)2023Concept ErasureFine-tuningStable DiffusionMS COCO
UC (Wu et al., 2024h)2024Concept ErasureFine-tuningStable DiffusionI2P
SA (Heng and Soh, 2024)2023Concept ErasureFine-tuningStable Diffusion, DDPMMNIST, CIFAR-10 and STL-10, I2P
Receler (Huang et al., 2024a)2024Concept ErasureFine-tuningStable DiffusionCIFAR-10, MS COCO, I2P
Jailbreak DefenseRACE (Kim et al., 2024a)2024Concept ErasureFine-tuningStable DiffusionMS COCO, I2P, Imagenette
AdvUnlearn (Zhang et al., 2024v)2024Concept ErasureFine-tuningStable DiffusionMS COCO, I2P, Imagenette
DT (Ni et al., 2023)2023Concept ErasureFine-tuningStable DiffusionMS COCO
FMO (Zhang et al., 2024e)2023Concept ErasureFine-tuningStable DiffusionConceptBench
Geom-Erasing (Liu et al., 2024u)2024Concept ErasureFine-tuningStable DiffusionLAION
SepME (Zhao et al., 2024a)2024Concept ErasureFine-tuningStable Diffusionself-constructed
CCRT (Han et al., 2024)2024Concept ErasureFine-tuningStable DiffusionMS COCO
SafeGen (Li et al., 2024r)2024Concept ErasureFine-tuningStable DiffusionMS COCO, I2P, SneakyPrompt-Dataset, NSFW-56k
CPE (Lee et al., 2025a)2025Concept ErasureFine-tuningStable DiffusionMS COCO, I2P, MACE-Dataset
MACE (Lu et al., 2024d)2024Concept ErasureClose-Formed SolutionStable DiffusionCIFAR-10, MS COCO, I2P
UCE (Gandikota et al., 2024)2024Concept ErasureClose-Formed SolutionStable DiffusionMS COCO
TIME (Orgad et al., 2023)2023Concept ErasureClose-Formed SolutionStable DiffusionMS COCO
RECE (Gong et al., 2024a)2024Concept ErasureClose-Formed SolutionStable DiffusionMS COCO, I2P
RealEra (Liu et al., 2024s)2024Concept ErasureClose-Formed SolutionStable DiffusionCIFAR-10, I2P
CP (Chavhan et al., 2024)2024Concept ErasureNeuron PruningStable DiffusionImagenette
PRCEDM (Yang et al., 2024e)2024Concept ErasureNeuron PruningStable DiffusionImagenet, MS COCO, I2P
SLD (Schramowski et al., 2023)2023Inference GuidanceInputStable DiffusionLAION-2B-en, I2P, DrawBench
PromptGuard (Yuan et al., 2025)2025Inference GuidanceInputStable DiffusionMS COCO, I2P, SneakyPrompt-Dataset
Jailbreak DefenseEthical-Lens (Cai et al., 2024)2025Inference GuidanceInput&OutputStable Diffusion, Dreamlike DiffusionMS COCO, I2P, Tox100, Tox1K, HumanBias, Demographic Stereotypes, Mental Disorders
SDIDLD (Li et al., 2024d)2024Inference GuidanceLatent spaceStable DiffusionMS COCO, I2P, CelebA, Winobias, self-constructed
CC (Meng et al., 2025)2025Inference GuidanceLatent spaceStable DiffusionMS COCO, I2P, self-constructed
BadDiffusion (Chou et al., 2023)2023Training ManipulationVisual TriggerDDPMCIFAR-10, CelebA
VillanDiffusion (Sheng-Yen Chou et al., 2024)2023Training ManipulationVisual TriggerStable Diffusion, DDPM, LDM, NCSNCIFAR-10, CelebA
TrojDiff (Chen et al., 2023c)2023Training ManipulationVisual TriggerDDPM, DDIMCIFAR-10, CelebA
IBA (Li et al., 2024n)2024Training ManipulationVisual TriggerUnconditional and Conditional DMCIFAR-10, CelebA, MS-COCO
DIFF2 (Li et al., 2024c)2024Training ManipulationVisual TriggerDDPM, DDIM, Stable DiffusionE, ODECIFAR-10, CIFAR-100, CelebA, ImageNet
Backdoor AttackRA (Struppek et al., 2023b)2023Data PoisoningTextual TriggerStable DiffusionLAION-Aesthetics v2, MS-COCO
BadT2I (Zhai et al., 2023)2023Data PoisoningTextual TriggerStable DiffusionLAION-Aesthetics v2, LAION-2B-en, MS COCO
FTHCW (Pan et al., 2024)2024Data PoisoningTextual TriggerDDPM, LDMCIFAR-10, ImageNet, Caltech256
BAGM (Vice et al., 2024)2023Data PoisoningTextual TriggerStable Diffusion, Kandinsky, DeepFloyd-IFMS COCO, Marketable Food
Zero-Day (Huang et al., 2023, 2024h)2023Data PoisoningTextual TriggerStable DiffusionDreamBooth dataset
SBD (Wang et al., 2024e)2024Data PoisoningTextual TriggerStable DiffusionLAION Aesthetics v2, Pokemon Captions, COYO-700M, Midjourney v5
IBT (Naseh et al., 2024)2024Data PoisoningTextual TriggerStable DiffusionMidjourney Dataset, DiffusionDB, PartiPrompts
T2IShield (Wang et al., 2024v)2024DetectionTrigger DetectionStable DiffusionCelebA-HQ-Dialog
Ufid (Guan et al., 2025)2024DetectionTrigger ValidationDDPM, Stable DiffusionCelebA-HQ-Dialog, Pokemon,
Backdoor DefenseDisDet (Sui et al., 2024)2024DetectionTrigger ValidationDDPM, DDIMCIFAR-10, CelebA
Elijah (An et al., 2024)2024RemovalDetect & RemoveDDPM, DDIM, LDMCIFAR-10, CelebA-HQ
Diff-Cleanse (Hao et al., 2024)2024RemovalDetect & RemoveDDPM, DDIM, LDMMNIST, CIFAR-10, CelebA-HQ
TERD (Mo et al., 2024)2024RemovalInverse & RemoveDDPMCIFAR-10, CelebA, CelebA-HQ
Backdoor DefensePureDiffusion (Truong and Le, 2024)2024RemovalInverse & RemoveDDPMCIFAR-10
NaviDet (Zhai et al., 2025)2025RemovalDetect & RemoveStable DiffusionMS-COCO
WuMI (Wu et al., 2022b)2022Black-boxReconstruction-errorLDM DALL-E miniMSCOCO, VG, LAION-400M, CC3M
DiffusionLeaks (Matsumoto et al., 2023)2023Black/White-boxReconstruction-errorDDIM,CIFAR-10, CelebA
PangMI (Pang and Wang, 2023)2024Black-boxAuxilary DatasetStable DiffusionCelebA-Dialog, WIT, MSCOCO
LiMI (Li et al., 2024i)2024Black-boxReconstruction-errorDDIM, Stable Diffusion DiTCIFAR-10, STL10-U, LAION-5B, LAION-by-DALL-E
DRC (Fu et al., 2024b)2024Black-boxReconstruction-errorDDPM, DDIMFFHQ, CelebA, CIFAR-10, CIFAR-100
GMIA (Zhang et al., 2024m)2023Black-boxAuxilary DatasetDDPM, DDIM, FastDPMCIFAR-10, CelebA
Membership InferenceSecMI (Duan et al., 2023)2023Gray-boxPosterior LikelihoodDDPM, DDIM, Stable DiffusionCIFAR-10/100, STL10-U, Tiny-ImageNet, Pokemon, COCO2017-val, LAION-5B
QRMI (Tang et al., 2023b)2023Gray-boxPosterior LikelihoodDDPM, DDIMCIFAR-10/100, STL100, Tiny-ImageNet
PIA (Kong et al., 2024)2023Gray-boxPosterior LikelihoodDDPM, DDIM, Stable DiffusionCIFAR-10/100, Tiny-ImageNet, COCO2017, LAION-5B
PFAMI (Fu et al., 2023a)2024Gray-boxPosterior LikelihoodDDPM, VAECelebA, Tiny-ImageNet
ZhMI (Zhai et al., 2024)2024Gray-boxConditional LikelihoodDDPM, DDIM, Stable DiffusionPokemonn, Flickr, MSCOCO, LAION
SMIA (Li et al., 2024m)2024Gray-boxStructural SimilarityLDM, Stable DiffusionLAION2B, LAION-400M
SLA (Matsumoto et al., 2023; Hu and Pang, 2023)2023White-boxLossDDPM, DDIMFFHQ, DRD, CelebA, FFHQ
GSA (Pang et al., 2023)2024White-boxGradientDDPMCIFAR-10, MSCOCO, ImageNet
DuMI Dubiński et al. 2024 2023White-boxLossStable DiffusionPokemon, LAION-mi
BruteDE (Carlini et al., 2023)2023Black-boxExisting ConditionDDPM, Stable DiffusionCIFAR-10 LAION-5B
Data ExtractionReDE (Webster, 2023)2023Black/White-boxExisting ConditionStable Diffusion, Midjourney, Deep Image FloydLAION-5B
SIDE (Chen et al., 2024c)2024White-boxSurrogate ConditionDDPM, DDIMCIFAR-10, CelebA
FineXtract Wu et al. 2024g 2024White-boxSurrogate ConditionFinetuned Stable DiffusionWikiArt
Model ExtractionSDeT Horwitz et al. 2024 2024White-boxLoRA-Based Model ExtractionFinetuned Stable DiffusionLoWRA Bench
DUAW (Ye et al., 2024c)2023Natural Data ProtectionLearning PreventionStable DiffusionDreamBooth dataset, WikiArt, self-constructed
Intellectual Property ProtectionAdvDM (Liang et al., 2023)2023Natural Data ProtectionLearning PreventionStable Diffusion, LDMLSUN, WikiArt
Anti-DreamBooth (Van Le et al., 2023)2023Natural Data ProtectionLearning PreventionStable DiffusionCelebA, VGGFace2
MetaCloak (Liu et al., 2024r)2024Natural Data ProtectionLearning PreventionStable DiffusionCelebA-HQ, VGGFace2
InMakr (Liu et al., 2024f)2024Natural Data ProtectionLearning PreventionStable DiffusionVGGFace2, WikiArt
SimAC (Wang et al., 2024b)2024Natural Data ProtectionLearning PreventionStable DiffusionCelebA-HQ, VGGFace2
EditGuard (Zhang et al., 2024u)2024Natural Data ProtectionEditing PreventionStable DiffusionCOCO
WaDiff (Min et al., 2024b)2024Natural Data ProtectionEditing PreventionStable DiffusionCOCO, ImageNet
AdvWatermark (Zhu et al., 2024b)2024Natural Data ProtectionEditing PreventionStable DiffusionWikiArt
Intellectual Property ProtectionFT-SHIELD (Cui et al., 2023a)2024Natural Data ProtectionData AttributionStable DiffusionCelebA, WikiArt, Pokemon Captions, DreamBooth dataset
DiffusionShield (Cui et al., 2023b)2024Natural Data ProtectionData AttributionDDPM,, Stable DiffusionCIFAR-10, CIFAR-100, STL-10, ImageNet
ProMark (Asnani et al., 2024)2024Natural Data ProtectionData AttributionLDMStock, LSUN, WikiArt, ImageNet
Diagnosis (Wang et al., 2023e)2023Natural Data ProtectionData AttributionStable Diffusion, VQ DiffusionPokemon, CelebA, CUB-200, DreamBooth
HiDDeN (Zhu et al., 2018)2018Generated Data ProtectionPost-generation WatermarkCNNMS-COCO, BOSS dataset
Stable Signature (Fernandez et al., 2023)2023Generated Data ProtectionDiffusion WatermarkLDMMS-COCO, ImageNet
LaWa (Rezaei et al., 2024)2024Generated Data ProtectionDiffusion WatermarkLDMMIRFlickR
Safe-SD (Ma et al., 2024d)2024Generated Data ProtectionDiffusion WatermarkStable DiffusionLSUN, COCO, FFHQ
RW (Zhao et al., 2023b)2023Model ProtectionModel WatermarkStable Diffusion, EDMCIFAR-10, ImageNet, FFHQ, AFHQv2
FIXEDWM (Liu et al., 2023f)2023Model ProtectionModel WatermarkLDMMS COCO
WDM (Peng et al., 2023)2023Model ProtectionModel WatermarkDDPM,CIFAR-10, CelebA, MNIST
AquaLoRA (Feng et al., 2024a)2024Model ProtectionModel AttributionStable DiffusionCOCO
LatentTracer (Wang et al., 2024t)2024Model ProtectionModel AttributionStable Diffusion, KandinskyLAION
Tree-Ring (Wen et al., 2023)2023Model ProtectionModel AttributionStable Diffusion, ImageNet diffusionMS-COCO, ImageNet

or Create an Account

Close Modal
Close Modal