Fabricated Identifier Hallucinations in Tool-Calling Agents
When AI agents invent fake reference IDs, systems execute them without question.
Tomás Vreeland
Staff Writer
Tomás Vreeland began his career writing about scientific misinformation for regional press outlets before narrowing his focus to the credibility failures of large language models and autonomous agents. He brings a skeptic's eye to vendor transparency claims and has developed a beat around red-teaming disclosures and hallucination rate reporting standards.
1 story
When AI agents invent fake reference IDs, systems execute them without question.