Which kinds of work people were ready to share with AI

In the 2023 survey, 79% were comfortable using AI for analytical work, 76% for administrative work and 73% for creative work. This is an early snapshot of willingness, not a statement about current adoption.

Keeping the year visible is especially useful when discussing changing technology. A past result can show what respondents anticipated at that time. It should not silently become a present tense claim about how people use tools today.

Citation's interpretation is to separate interest in a task from the ability to complete it responsibly. Someone may be comfortable asking for a summary while still needing help checking whether it represents the source fairly. Another person may know how to verify the result but have little reason to use the tool.

A practical introduction can use a small assignment with a clear finish. For example, ask someone to compare two short source passages, then explain where the generated summary preserved or lost a distinction. This connects the tool with a task and an observable check.

Broad categories conceal different levels of responsibility. Suggesting a title and preparing a published argument both involve creative work, but they need different review. Sorting a table and drawing a causal conclusion both involve analysis, but they do not ask the same thing of the evidence.

Define the exact activity before interpreting readiness. Then decide which inputs are appropriate, which parts require review and who owns the final decision. The chart offers broad attitudes; a useful workflow needs those more concrete choices.

The 2023 Work Trend Index surveyed 31,000 employed or self employed workers across 31 markets between February 1 and March 14. The full study includes different work settings. The three task measures can overlap and the public narrative does not state a separate response count for each item.

The original report provides the published figures and question wording. The chart and downloadable data use the same values; labels are shortened for readability.