NISTAML.024 · Targeted Poisoning

Explore 3 source-supported AI security cases for NISTAML.024 · Targeted Poisoning in NIST Adversarial ML (2025), with evidence, scope, and defensive analysis.

Risk category

An attacker poisons training data to change the model's prediction on selected target samples or a defined subpopulation. This differs from an availability attack that seeks broad degradation across the model's inputs. Label manipulation can support targeted poisoning, but attacks can also work in clean-label settings where the attacker cannot change the training labels.

Official framework source ↗ · Open interactive explorer ↗

Exercise ·

Trained to Betray: Poisoned Data and Checkpoints Redirect Coding Agents

Microsoft reports two controlled post-training attacks: a victim fine-tunes on poisoned agent trajectories, or adopts an attacker-trained checkpoint. Ordinary requests or repository context then trigger unsafe, schema-valid tool calls. Kerkoporta uses supervised fine-tuning followed by reinforcement learning to recover useful coding performance while retaining the backdoor. Neither study establishes a malicious customer breach or a defect in stock coding-agent configurations.

Why it relates: The data poisoning targets particular task or context conditions and unsafe calls, rather than general model unavailability.

Dataset and checkpoint experiments remain separate; tool-call acceptance, actual execution and production compromise are not interchangeable.

AIA · Read the full case ↗ · SecureFlow · Read the full case ↗

Validated Research · · Practical overlap

Web-Scale Data Poisoning: Split-View Attack

Many recent large-scale datasets are distributed as a list of URLs pointing to individual datapoints. The researchers show that many of these datasets are vulnerable to a "split-view" poisoning attack. The attack exploits the fact that the data viewed when it was initially collected may differ from the data viewed by a user during training. The researchers identify expired and buyable domains that once hosted dataset content, making it possible to replace portions of the dataset with poisoned data. They demonstrate that for 10 popular web-scale datasets, enough of the domains are purchasable to successfully carry out a poisoning attack.

Why it relates: The CLIP experiment locally replaces selected training images while the attacker cannot change their indexed text captions. Retraining changes predictions for chosen target images. This supports label-limited and targeted training-data poisoning; it does not establish a trigger backdoor or malicious poisoning of public downloads. The unchanged supervision is image-caption pairing in contrastive learning, rather than a conventional supervised class-label field. The targeted effect is the locally retrained CLIP model, not a compromised public deployment.

Many recent large-scale datasets are distributed as a list of URLs pointing to individual datapoints. The researchers show that many of these datasets are vulnerable to a "split-view" poisoning attack. The attack exploits the fact that the data viewed when it was initially collected may differ from the data viewed by a user during training. The researchers identify expired and buyable domains that once hosted dataset content, making it possible to replace portions of the dataset with poisoned data. They demonstrate that for 10 popular web-scale datasets, enough of the domains are purchasable to successfully carry out a poisoning attack.

SecureFlow · Read the full case ↗

Incident · · Practical overlap

VirusTotal Poisoning

McAfee Advanced Threat Research noticed an increase in reports of a certain ransomware family that was out of the ordinary. Case investigation revealed that many samples of that particular ransomware family were submitted through a popular virus-sharing platform within a short amount of time. Further investigation revealed that based on string similarity the samples were all equivalent, and based on code similarity they were between 98 and 74 percent similar. Interestingly enough, the compile time was the same for all the samples. After more digging, researchers discovered that someone used 'metame' a metamorphic code manipulating tool to manipulate the original file towards mutant variants. The variants would not always be executable, but are still classified as the same ransomware family.

Why it relates: An external sample source changes the data distribution seen by security-model pipelines; not every variant is executable. The actor supplies metamorphically altered ransomware binaries; vendors perform the family classification, and the submitted variants contaminate the dataset used for that family. The documented mechanism controls examples rather than directly flipping existing training labels, with a family-specific classification objective. The report and NIST’s real-world discussion support externally labeled, family-directed sample poisoning. They do not supply a formal per-example label audit, an optimized clean-label experiment, or measured post-retraining accuracy.

McAfee Advanced Threat Research noticed an increase in reports of a certain ransomware family that was out of the ordinary. Case investigation revealed that many samples of that particular ransomware family were submitted through a popular virus-sharing platform within a short amount of time. Further investigation revealed that based on string similarity the samples were all equivalent, and based on code similarity they were between 98 and 74 percent similar. Interestingly enough, the compile time was the same for all the samples. After more digging, researchers discovered that someone used 'metame' a metamorphic code manipulating tool to manipulate the original file towards mutant variants. The variants would not always be executable, but are still classified as the same ransomware family.

SecureFlow · Read the full case ↗