Source: [GWI Zeitgeist](https://www.gwi.com/zeitgeist) (full disclosure, I work for the company that pulled this research together)
Tools used: Python w/ WordCloud library
I’ve been tracking how people feel about AI for a while, but normally in the form of quantitative surveys (which tell you “x% of people are worried/excited about AI”). For our latest round we ran a test of an open-ended question, where respondents could write whatever they liked, and the wordcloud here displays their most commonly used words in those responses. NB: this survey was fielded among ~4,000 respondents in the US and UK only.
Even with a slightly different methodology, it reflects a pattern we’ve seen for a number of years (even before ChatGPT came along) – people are really, really ambivalent about AI. They’re excited about it, but worried in equal (or greater) measure.
Would be nice to see something like this done with data gathered exclusively before 2022, and compare with this one.
As someone who works in data, I find the term ‘AI’ really annoying because it’s way too vague.
Sometimes people use it to refer to LLMs like ChatGpt or Gemini or Claude or whatever, or neural nets like Google Alphafold or whatever that have specific applications.
Sometimes they refer to it as some miracle future technology that doesn’t exist yet.
And sometimes they use it to mean automation of jobs, which is something that has been happening since the Luddite revolt in the industrial revolution in the 1800s.
And all these things are completely different.
Would be interesting to see this with sentiment analysis on the words so the colour coding indicated positive/negative. Am I right in thinking that currently it’s just coloured for aesthetic reasons?
![[OC] Words people most frequently use to describe how they feel about AI](https://www.europesays.com/wp-content/uploads/2025/01/5rsx4cnrgjee1-1920x1024.png)
4 comments
Source: [GWI Zeitgeist](https://www.gwi.com/zeitgeist) (full disclosure, I work for the company that pulled this research together)
Tools used: Python w/ WordCloud library
I’ve been tracking how people feel about AI for a while, but normally in the form of quantitative surveys (which tell you “x% of people are worried/excited about AI”). For our latest round we ran a test of an open-ended question, where respondents could write whatever they liked, and the wordcloud here displays their most commonly used words in those responses. NB: this survey was fielded among ~4,000 respondents in the US and UK only.
Even with a slightly different methodology, it reflects a pattern we’ve seen for a number of years (even before ChatGPT came along) – people are really, really ambivalent about AI. They’re excited about it, but worried in equal (or greater) measure.
Would be nice to see something like this done with data gathered exclusively before 2022, and compare with this one.
As someone who works in data, I find the term ‘AI’ really annoying because it’s way too vague.
Sometimes people use it to refer to LLMs like ChatGpt or Gemini or Claude or whatever, or neural nets like Google Alphafold or whatever that have specific applications.
Sometimes they refer to it as some miracle future technology that doesn’t exist yet.
And sometimes they use it to mean automation of jobs, which is something that has been happening since the Luddite revolt in the industrial revolution in the 1800s.
And all these things are completely different.
Would be interesting to see this with sentiment analysis on the words so the colour coding indicated positive/negative. Am I right in thinking that currently it’s just coloured for aesthetic reasons?
Comments are closed.