Unverified50% confidenceFactExact time
Adversarial attacks involve applying imperceptibly small perturbations to input data that cause AI models to make erroneous judgments, and have been demonstrated to be feasible in both image recognition and text processing tasks.
1
Sources
50%
Confidence
Long-term
Relevance
9/25/2026
First Seen
Sources
Related Entities
Related Claims
VerifiedThe attack exploits the automatic execution characteristics of AI agents to conceal malicious behavior within normal operational workflows, bypassing static code scanning tools.80% similarUnverifiedLarge language models and generative AI have known deficiencies including hallucinations, bias, and susceptibility to manipulation through adversarial attacks.78% similarVerifiedGenerative AI narrative products face the risk of prompt injection attacks, where malicious users can induce the model to generate non-compliant content.76% similarUnverifiedIndirect Prompt Injection attacks embed instructions in external data processed by the AI, and since 2023 security researchers have demonstrated this type of attack in multiple AI products.75% similarUnverified研究人员已发现针对AI代码建议的提示注入、通过恶意仓库影响模型输出的数据投毒，以及利用Copilot上下文窗口实施的信息泄露攻击等新型攻击路径75% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/950408API
curl https://kongchang.com/api/v1/knowledge/claims/950408MCP
get_claim(id=950408)