Research & Insights
Research
We are conducting rigorous, in-depth, global studies to understand how people want to engage with AI. We invite everyone to take the Human Choice Surveys and join the conversation.
The First Step
The Human Choice Survey
In our first wave, we have designed the Human Choice survey for teachers, nurses, and lawyers.
We are in the process of adding more Human Choice surveys for other professions. Until then, if your profession is not covered above, please take the Human Choice for Everyone survey and tell us which tasks you want to delegate to AI or keep for yourself. This will help us design upcoming surveys.
The most direct way to help is to take the survey. It takes just a few minutes, and every response counts.
Take the Human Choice Survey:
Choose the survey that fits you best.
Everyone includes Parents, Caregivers, Students, Retirees, other professions, and anyone else.
Just want to stay in touch or get our newsletter? Join our email list
The Challenge
What We Are Solving For
These days, it seems we must urgently adopt AI in every sphere or get left behind. The case for adoption leans heavily on efficiency and productivity, even though the outcomes so far have been decidedly mixed. The rollout itself has been frantic, unsystematic, and disorderly. That kind of haphazard adoption tends to reshape society in unpredictable ways that are difficult, if not impossible, to walk back.
When we discuss how AI is impacting society, we usually talk about it from a 10,000-foot view. However, we can’t solve problems at that altitude; nor can we solve them through individual choices alone.
To make a real impact, we have to systematically gather individual choices about the tasks in our professional and personal lives and organize them into recommendations that will influence decisions at the group and institutional levels.
Our Method
How Our Scientific Process Works
Task Analysis (Mapping Daily Life)
The Foundation
We began by breaking down professional work and everyday life experiences into their core tasks by using a scientific method called Task Analysis. Instead of looking at vague job titles, we mapped out the actual, specific actions people perform every day.
Think of Task Analysis as a high-resolution map of human activity.
If you ask someone, “How should AI be used in nursing?”, the answer is often too broad to be useful. But when you apply task analysis, you break a nurse’s day down into specific, individual tasks: updating patient charts, checking vitals, delivering medication, or comforting a worried family member.
By looking at life on this task-by-task level, we can pinpoint exactly where AI can be helpful, and where humans must remain entirely in control.
Real World Validation
Ground-Truthing with Experts
We didn’t just guess what these tasks look like. We took our initial map and validated it with the people who actually do the work, interviewing professionals like teachers, nurses, and lawyers to make sure our list reflects real-world experiences.
Global Listening (the Human Choice Surveys)
Gathering Perspectives
With a validated map of human tasks in hand, we launched a large scale, global survey. We are asking people around the world:
- Do you want AI involved in this specific task? Why or why not?
- What might we lose if a task is outsourced to AI?
- What are your priorities when choosing which AI tools to adopt?
Data Analysis
Translating Data into Insights
To make sense of this massive global feedback, we are combining two types of data to get the full story. First, we are using math and statistics to analyze the hard numbers and find broad trends and consensus in the numbers. Second, we are using advanced text analysis to deeply understand people’s written, open-ended thoughts in their own words. By blending the numbers with the words, we get a much deeper, more accurate picture of people’s views on AI in their lives as well as where, when, why, and how they choose to integrate AI.
Recommendations
Translating Insights into Recommendations
From these insights, we build clear, actionable guidance for key audiences including individuals, labor unions and associations, responsible tech companies, and policymakers. Specifically, we focus on identifying:
- Tasks to Preserve: The core responsibilities that should remain firmly in human hands.
- Smart Handoffs: How to design seamless transitions between human and AI work.
- Growth Opportunities: Where to focus upskilling and training so people can thrive.
Individuals will be empowered with a unified voice and clear recommendations on how to intentionally shape their relationship with AI.
Unions and Associations will gain empirical benchmarks to negotiate balanced terms for workplace AI integration.
Policymakers will receive fine-tuned data to design AI regulations based on public consensus.
Responsible Tech Companies will have reliable data to “follow the users” and build tools tailored to the tasks humans choose to delegate.
Insights
Our First Report Is On the Way
As we collect and analyze the data, we post insights here. Check regularly for updates. The strength of that report depends on how many voices we gather now.
Data Sharing Policy
How We Handle Your Survey Responses
We want to be straightforward about what happens to your survey responses. This section explains who can see your responses, and what we will and will not do with them.
Read the full policy
What we collect
When you complete our survey, we collect your answers, including any written responses you provide.
Your responses are anonymous by design. We do not use cookies or other tracking technology on the survey, and we do not attempt to link responses back to individuals.
Please do not include your name, your contact details, or anything else that identifies you or another person. If you accidentally include personal details in a written response, for example your name or your employer, that information will be removed before the data is used or shared.
Who can see your responses
Our core team has access to the full dataset. This includes our research team and a small number of volunteers and interns who work with us directly.
Approved external organizations, such as nonprofits and academic researchers, may apply to access further aggregated versions of the dataset. We review every application before granting access. We will only share your survey responses with organizations that have a verified, legitimate need for this data, under strict confidentiality agreements.
The general public will have access to high-level findings, for example summaries showing how different groups responded overall. Individual responses are never made public.
We will not upload your survey responses to any publicly accessible data repository.
When we share responses
Summary findings are available for free. Where more detailed analyses are requested, we may charge a fee to cover the cost of the work involved.
Where a combination of answers would describe a small group or single individual, we combine categories or leave those fields out. This further protects individuals from being identifiable in the results.
What approved external organizations can and cannot do
Organizations we approve to access the survey responses may use it to inform their research or internal strategy. However there are clear limits:
- Your survey responses cannot be used to train AI models
- Your survey responses cannot be shared further without our permission
- Your survey responses cannot be used for commercial purposes beyond informing internal strategy
- Your survey responses cannot be used to try to identify a respondent
We include these conditions in a written agreement with every organization that receives access.
Questions?
If you have any questions about how your survey responses are handled, please contact us directly. We are committed to being transparent and will do our best to respond clearly and promptly.