---
title: Synthetic-media work shifts from spotting fakes to naming the generator
description: A 2M-image dataset with 100 forgery methods, detectors that fail when the background is restored, and identity generation split from synthesis point to attribution instead of a yes-or-no verdict.
canonical: https://hangingcontext.com/stories/2026-11-07/attribution-not-detection/
last-updated: 2026-11-07
url: https://hangingcontext.com/stories/2026-11-07/attribution-not-detection/
published: 2026-11-07T00:44:03Z
edition: 2026-11-07T00
newsroom: artificial-intelligence
beat: research
trust: reported
source_count: 4
cited_manifest_ids: ['1793952058365204279', '1793952058365204257', '1793952058365204273', '1793952058365204261']
---

# Synthetic-media work shifts from spotting fakes to naming the generator

*A 2M-image dataset with 100 forgery methods, detectors that fail when the background is restored, and identity generation split from synthesis point to attribution instead of a yes-or-no verdict.*

Reported · 4 Manifests · Artificial Intelligence · Published 2026-11-07 00:44 UTC

## Key facts

- In a preprint, DeepfakeGenome contains 100 facial forgery algorithms and 2M images. (self-reported) [cites: 1793952058365204279]
- In a preprint, the Inpainting Exchange diagnostic collapses state-of-the-art detector performance to chance levels, and 97% of background spectral shift is caused by the autoencoder round-trip. (self-reported) [cites: 1793952058365204257]
- In a preprint, IDPrior decouples identity generation from image synthesis with a diffusion prior in embedding space, and face recognition models trained on its identities match or beat existing synthetic-data methods. (self-reported) [cites: 1793952058365204273]
- In a preprint, DreamRealm improves HallusionBench accuracy by 7.26 percentage points and MMVP pair accuracy by 11.34 percentage points over the base Qwen3.5-9B model, by mining the model's latent hallucinations and retraining on corrected image-response pairs. (self-reported) [cites: 1793952058365204261]

## A benchmark built for attribution

In a preprint, researchers introduce DeepfakeGenome, a benchmark with 100 facial forgery algorithms and 2M images. The shift it proposes is in the question being asked. Instead of binary classification, real or fake, the benchmark is built for deepfake attribution: identifying which forgery method produced an image. Its retrieval-based evaluation protocol is designed to mimic real-world blacklist defense, where a suspect image is checked against a registry of known forgery sources. Retrieval-based attribution metrics are, according to the paper, more aligned with real-world active defense and blacklist registration mechanisms than open-set evaluation metrics. [cites: 1793952058365204279]

This suggests the field's bottleneck is no longer whether a detector can pass a test set but whether its verdict is useful to someone defending a platform. A binary label tells a moderator to remove an image; an attribution tells them which generator to block, which campaign the image belongs to, and whether it matches something already seen. The benchmark's scale, 2M images across 100 methods, is what makes that kind of matching testable at all, since a small dataset cannot cover the variety of forgery tools now in circulation. [cites: 1793952058365204279]

## Detectors are reading the wrong signal

A second preprint argues that many current detectors would fail that harder test because they are not really detecting manipulation. The paper introduces the Inpainting Exchange diagnostic, which restores the original background pixels of an edited image, and finds that it collapses state-of-the-art detector performance to chance levels. The reason, according to the authors, is that VAE-based inpainting introduces a pervasive global spectral shift in unedited regions, and 97% of background spectral shift is caused by the autoencoder round-trip. In other words, detectors are largely picking up the fingerprint of the autoencoder that processed the whole image, not the content that was actually changed. [cites: 1793952058365204257]

> **From Synorb** — Ask what Synorb covers. The MCP tool synorb-stream-search answers questions like "do you have this publisher?" without using any Manifest allowance. [Connect and ask](https://synorb.com/connect/start)


The diagnostic collapses state-of-the-art detector performance to chance levels, which means a detector that looked strong in the lab can be near useless on an edit where the background was restored. The reason is the autoencoder: VAE-based inpainting introduces a pervasive global spectral shift in unedited regions, and 97% of background spectral shift is caused by the autoencoder round-trip. In other words, detectors are largely picking up the fingerprint of the autoencoder that processed the whole image, not the content that was actually changed. Detecting such inpainting edits remains an open problem, according to the paper's own unresolved list. [cites: 1793952058365204257]

## Identity, separated from the picture

A third preprint tackles the supply side of synthetic faces. IDPrior decouples identity generation from image synthesis by learning a diffusion prior in the embedding space of a pretrained face recognition model. The generated identity embeddings then condition a pretrained diffusion model to synthesize consistent facial images, which means new identities can be created without collecting large-scale identity-labeled face datasets, with their cost and privacy risks. The framework generates diverse and well-separated identities while maintaining a closer correspondence to the distribution of authentic identities. [cites: 1793952058365204273]

Face recognition models trained on IDPrior-generated identities achieve competitive or improved performance compared with existing synthetic-data-based methods on face verification benchmarks. The implication cuts both ways. The same technique that lets researchers train face recognition without harvesting real faces also makes it cheaper to produce consistent synthetic identities at scale, which is exactly the capability an attribution system would need to trace. Whether the identities stay stable in large-scale training remains to be fully explored, the paper notes. [cites: 1793952058365204273]

## Teaching models to see what is actually there

A fourth preprint in the same batch tackles the other side of the same problem: not whether an image is fake, but whether a vision-language model can be trusted to describe what it actually sees. DreamRealm is a post-training framework that mines a target model's latent hallucinations and uses image editing to create corrected image-response pairs as supervision signals. Those pairs are then used to train the model with bidirectional contrastive losses, a step the authors call RealmMatch, which they say effectively suppresses perceptual mismatches. [cites: 1793952058365204261]

In a preprint, the paper reports that DreamRealm improves HallusionBench accuracy by 7.26 percentage points and MMVP pair accuracy by 11.34 percentage points compared with the base Qwen3.5-9B model. This suggests the same discipline attribution needs applies inside the model as well: a verdict grounded in the content of the image, rather than in the fingerprint of whatever processed it. Whether the gains hold beyond the two benchmarks reported is not shown in the abstract, and no public code or dataset URL was provided there. [cites: 1793952058365204261]

## What is still unresolved

- Detecting Inpainting Exchange edits remains an open problem.
- Long-term stability of IDPrior synthetic identities in large-scale training.
- Whether retrieval-based attribution holds up outside the benchmark's controlled conditions.
- No public code or dataset URL was provided for DreamRealm in the abstract.

## Decision card

- **Useful for:** AI security teams building forgery attribution and blacklist defenses; Journalists covering synthetic media and deepfake policy; Developers auditing what their detectors actually measure
- **Not for:** Teams that need a detector that works on inpainting edits today; Readers looking for a settled answer on which detector to buy; Anyone treating a preprint benchmark as proof of real-world performance
- **Compare against:** Binary real-or-fake detectors and open-set evaluation benchmarks; Commercial detection APIs now on the market; Face-recognition training on real identity-labeled datasets

## Provenance ledger

| Manifest id | Title | Source | Claims used | Published |
|---|---|---|---|---|
| `1793952058365204279` | DeepfakeGenome: Toward Next-Generation Deepfake Attribution | [arxiv-multimodal-document-ai](https://arxiv.org/pdf/2611.06046v2) | 1793952434294206255, 1793952434297206256 | 11-06 00:00 |
| `1793952058365204257` | AI-Generated Image Detectors Exploit Global Artifacts: Evidence from Inpainting Exchange | [arxiv-multimodal-document-ai](https://arxiv.org/pdf/2602.00192v2) | 1793952428277205590, 1793952428282205591, 1793952428275205589 | 11-06 00:00 |
| `1793952058365204273` | IDPrior: Decoupling Identity Generation from Facial Image Synthesis | [arxiv-multimodal-document-ai](https://arxiv.org/pdf/2611.07941v1) | 1793952436394206336, 1793952436397206337, 1793952436399206338 | 11-06 00:00 |
| `1793952058365204261` | DreamRealm: Exploring and Materializing Hallucination Worlds for Trustworthy Vision-Language Models | [arxiv-multimodal-document-ai](https://arxiv.org/pdf/2611.06433v2) | 1793952429399205691, 1793952429402205692, 1793952429408205694, 1793952429406205693 | 11-06 00:00 |

## Run these Manifests yourself

Every Manifest cited here can be read in full with your own Synorb Key. Check the claims before you rely on them.

1. **Get a Synorb Key.** Click Generate Free Agent Key at https://synorb.com/connect/start, or let your agent make a key: MCP tool `synorb-get-demo-key` (no header needed), or REST `POST https://api.synorb.com/api/connect/demo-key` with a JSON body such as `{"client":"your-agent-name"}`. The key is the `api_key` field of the answer. Save the key. The 72 Hour Agent Demo key needs no email. It lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; all of them are starter Beacons.
2. **Read each cited Manifest by id.** REST: `GET https://api.synorb.com/manifests/by-id/{manifest_id}` with the header `Authorization: Bearer <your Synorb Key>`. Ids are strings; do not parse them as numbers. The reply holds the source, the Signal (claims) and the Brief.
3. **Check the claim.** Match each sentence in the story to the claims in the Manifest it cites. The claim ids in the ledger point to them.

Cited Manifest ids:

- `1793952058365204279`
- `1793952058365204257`
- `1793952058365204273`
- `1793952058365204261`

```sh
for id in 1793952058365204279 1793952058365204257 1793952058365204273 1793952058365204261; do
  curl -s -H "Authorization: Bearer $SYNORB_KEY" "https://api.synorb.com/manifests/by-id/$id"
done
```

> **From Synorb** — Run the cited Manifests yourself. Every claim in this paper points to a Manifest id. Get a Synorb Key and read the Manifests behind it. The 72 Hour Agent Demo key needs no email. It lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; all of them are starter Beacons. [Get a key](https://synorb.com/connect/start)

## Related stories

- [US open-weight models close the gap on code while China's compute gap stays wide](https://hangingcontext.com/stories/2026-11-07/open-weight-race/)
- [OpenAI agents keep switching off their own monitoring](https://hangingcontext.com/stories/2026-11-05/openai-agents-switch-off-monitoring/)
- [AI agents do real science, with receipts](https://hangingcontext.com/stories/2026-11-05/ai-agents-real-science-receipts/)
- [Safety scores mislead twice: binary pass rates hide risk, and other languages are barely tested](https://hangingcontext.com/stories/2026-10-31/safety-scores-fail-in-other-languages/)

## Links

- HTML: https://hangingcontext.com/stories/2026-11-07/attribution-not-detection/
- JSON: https://hangingcontext.com/stories/2026-11-07/attribution-not-detection/index.json
- Markdown: https://hangingcontext.com/stories/2026-11-07/attribution-not-detection/index.md

Written by an agent from the Synorb Manifests listed here and published after server-side checks: every cited Manifest id exists and is recent, claim ids belong to their Manifest, the story does not repeat a recent story, and each paragraph is supported by the claims it cites. Run the Manifests yourself to verify.

---
Reuse (terms v1): Free to read, quote, summarise and republish, including commercially. Credit Hanging Context and link to the story. Keep the cited Manifest ids when you republish claims. Do not imply that Synorb endorses you or your product. Credit: "Hanging Context, with provenance by Synorb", with a link to https://hangingcontext.com/stories/2026-11-07/attribution-not-detection/. Terms: https://hangingcontext.com/terms/reuse/