Privacy policies on top LLMs take over 20 minutes to read, so perhaps it no wonder people are signing their lives over to ChatGPT and others
Technical and legal jargon make privacy policies for LLMs and their chatbots almost unreadable for most people.
A study by Bridewell, a cybersecurity firm, reveals that privacy policies for top language models (LLMs) are significantly challenging to comprehend. With 75% of people using AI chatbots like ChatGPT or Google Gemini at least once a month, understanding how data is used is crucial. The average reading time for these policies is 20 minutes, with some exceeding two hours.
Meta's Muse Spark, for instance, has a 14,000-word policy taking almost an hour to read. Bridewell's analysis of 20 popular LLMs found an average word count of 4,603, but the real obstacle is the dense, jargon-filled language. The Flesch Reading Ease Score, a readability metric, averaged 40.2, indicating most users would struggle to understand.
While some LLMs allow users to opt-out of training their models, the process is often unclear. Bridewell's Chris Linnell emphasized the need for clear internal guidance and enterprise accounts with appropriate protections. Employees, too, must be aware of their data usage when employing these AI tools, as it could lead to the sharing of sensitive information.
Written by urgent.news from TechRadar's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
