English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The interactive use of large language models (LLMs) in AI assistants (atwork, home, etc.) introduces a new set of inference-time privacy risks: LLMsare fed different types of information from multiple sources in their inputsand are expected to reason about what to share in their outputs, for whatpurpose and with whom, within a given context. In this work, we draw attentionto the highly critical yet overlooked notion of contextual privacy by proposingConfAIde, a benchmark designed to identify critical weaknesses in the privacyreasoning capabilities of instruction-tuned LLMs. Our experiments show thateven the most capable models such as GPT-4 and ChatGPT reveal privateinformation in contexts that humans would not, 39% and 57% of the time,respectively. This leakage persists even when we employ privacy-inducingprompts or chain-of-thought reasoning. Our work underscores the immediate needto explore novel inference-time privacy-preserving approaches, based onreasoning and theory of mind.

Can LLMs Keep a Secret? Testing Privacy Implications of Language Models via Contextual Integrity Theory