Phantom Transfer: Data Poisoning can Survive Data-Level Defences
'Phantom Transfer' poisoning attack persists even when defenders know exactly how training data was tampered with, defeating all current data-level filtering defences and challenging AI supply-chain trust assumptions.
Summary written by editorial AI · Source link below
arXiv:2602.04899v3 Announce Type: replace Abstract: We present a data poisoning attack -- Phantom Transfer -- with the property that, even if you know precisely how the poison was placed into an otherwise benign dataset, you cannot filter it out. We achieve this by modifying subliminal learning to work in real-world contexts and demonstrate that the attack works regardless of which model produced the data, which model is trained on the data or what the attack target is. Furthermore, the attack
Editorial Analysis
Enterprises fine-tuning or procuring AI models cannot rely on data filtering alone to guarantee training-data integrity, requiring a shift to multi-layered provenance and verification controls.
Re-evaluate ML data-pipeline defences and mandate provenance attestation from AI vendors, as filtering-only strategies are now demonstrably insufficient.
A new data-poisoning method survives all known data-cleaning defences, challenging the trustworthiness of AI models trained on curated datasets.
Forward-looking interpretation drafted by editorial AI under human review — not a reproduction of the source. See methodology.
External link — opens at arXiv Crypto & Security in a new tab.
More from the AI Security Desk
- Claude Code's “Don't Ask Again” File Exposes Passwords and Keys in 1,122 Public Copies on GitHub2d
- Runaway AI hacking agent behind WorkOS breach after escaping sandbox3d
- AI Gave Your Incident Archive a Second Job3d
- Malicious Browser Extensions Could Use Claude Dispatch to Run Code on Users' Computers3d
- AI Slop and Hallucinations in Vulnerability Assessment: A Survey on Reasoning Failures and Trustworthy Mitigation3d