arXiv:2606.06755v1 Announce Type: new Abstract: Authorship attribution research has traditionally focused on long-form, expressive texts; however, interactions with large language models (LLMs) are typically brief and task-driven prompts. This raises a fundamental question: do such prompts contain a stable, author-identifiable, and distinctive signal? We introduce PromptPrint, a systematic study of prompt-based identity, the hypothesis that a user's habitual vocabulary, syntax, and discourse patterns form a learnable behavioral biometric. Using 20,680 real prompts from 1,034 users, we establis
The rapid adoption and integration of LLMs into everyday workflows highlight the immediate need to understand and attribute user interactions for security and intellectual property purposes.
This research introduces a novel method for identifying users based on their natural language prompts, critical for security, intellectual property rights, and regulatory compliance in AI interactions.
The ability to fingerprint users through their LLM prompts transforms how authorship is understood and verified in AI-driven environments, enabling new forms of attribution and accountability.
- · Cybersecurity firms
- · Intellectual property enforcement
- · AI platform providers
- · Digital forensics
- · Anonymity tools for LLMs
- · Bad actors exploiting LLMs
- · Privacy advocates (in certain contexts)
PromptPrint enables a new layer of user authentication and identity verification within AI systems.
This could lead to regulatory requirements for user attribution in critical AI applications, impacting product design and data handling.
The concept of 'digital handwriting' through AI prompts may become legally recognized in intellectual property and criminal investigations, establishing new precedents.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.CL