How do you imagine artificial intelligence (AI)? A marvel, a menace, or a mystery out of your control? At Digital Public Square, we believe that people should have the tools, and the space, to make sense of technology and how it affects their communities.
AImagine is a personality quiz that helps you explore how you imagine AI. Built in response to the Global Dialogues Challenge from The Collective Intelligence Project, the experience shows how your perspective compares to over 2,000 people from around the world.
It is essential that we can have more inclusive and meaningful conversations about AI. But, these conversations can be difficult. While the popularity of AI in research, technology, business, and policy has exploded, the term is notoriously slippery. “Artificial intelligence” can refer to everything from an algorithm that runs on a Game Boy to a language model that requires a warehouse of servers. Today, the focus is usually on generative technologies that we interact with through natural language.
Even when people share a basic understanding of AI technology, their different perspectives can create polarizing conversations. The social, environmental, and economic impacts and opportunities of AI have created rifts among both industry experts and the general public. AImagine’s questions help you explore your own perspectives, and presents a range of profiles so that you can think about those of others.
Take the quiz to discover your AImagine profile—and see how your perspective fits into the global dialogue. Or, read on to learn more about how we built the experience.
Thanks for reading Digital Public Square’s Newsletter! Subscribe for free to receive new posts and support our work.
How we made AImagine
Analylzing the data
The Global Dialogues datasets include waves of questions asked to respondents around the world (a bit over 1,000 respondents per wave). Some questions are multiple choice while others are open text. Our analysis focused on multiple choice questions that were asked across the two most recent waves.
First, the responses were converted into numbers (for example, “Disagree” = 1, “Unsure” = 2, “Agree” = 3). This helped us compare the different responses to each other. Using these numeric representations of each respondent’s data, we used principle component analysis and K-means clustering algorithms to identify four groups of respondents that had similar answers. If you head to the profiles page of the quiz, these clusters roughly correspond to the Believer, Skeptic, Power-user, and Doomer profiles.
Developing the profiles
Examining the responses of these different clusters helped us identify our key AI characteristics:
- Beneficial: how much you believe that AI is harmful versus beneficial to society
- Effective: how much you believe that AI is productive versus error-prone
- Agentic: how much you believe that AI can think or act independently versus only with human control
Believers and Power-Users both saw AI as effective and beneficial for society, while Skeptics and Doomers saw AI as harmful. Believers and Power-Users were differentiated based on their attitudes towards AI agency. Believers trusted AIs to operate autonomously and viewed them more as intelligent entities while Power-Users preferred to stay in control. Skeptics and Doomers differed both in their perceptions of how effective and agentic AI is, with Doomers seeing it as much more powerful than Skeptics.
