Our finite lives in an age of endless content | The Interflourishing Institute

How AI can help us care for human attention and make room for meaningful living

Every article, message and video asks for a little of someone’s life. Reading takes time. So does deciding whether something deserves to be read, checking whether it is true and working out what to do with it. That time comes from the same day in which a person works, cares for others, eats, rests and tries to find room for meaning and purpose.

At Interflourishing, our starting point is an ethical commitment: human attention deserves care because it is part of a finite life. AI should help us spend that life meaningfully.

The reference point here is generative AI as it is used in November 2026. Some systems can do more than respond to individual prompts. AI agents can plan a sequence of actions, use tools and adjust their approach without a person directing every step. Their behaviour is shaped by their training, instructions, available tools and the information they can access. AI developer Anthropic describes these capabilities and the continuing difficulty of ensuring that agents act as users intend. Anthropic, 2026.

AI can also generate starting ideas. In a 2024 experiment, people given access to AI-generated story ideas produced short stories that were rated as more creative, although their stories were also more similar to one another. This does not establish the quality or originality of every AI-generated idea, however. Doshi and Hauser, 2024.

The ability to generate ideas or carry out a task leaves an important responsibility with the people using the system. We still need to decide what purpose the work serves and whether the result deserves someone else’s attention. AI can help us explore those questions, but we remain responsible for what we choose to publish and ask others to engage with.

Generative AI can reduce the time it takes to draft writing or create images. This can make useful work possible. It also makes it easier to produce material without giving enough thought to the people who will receive it. Our ability to generate another article or report tells us little about whether someone needs it.

Consider a hypothetical team using AI to turn brief notes into a long report. The person preparing it finishes sooner. Their colleagues still need to read it, check its claims and identify the decisions it requires. If the report adds unnecessary explanation or errors, some of the effort saved by its author has become work for everyone else.

The same tool could help that team prepare a clear account of what happened, what needs deciding and where the supporting information can be found. The difference lies in the purpose and care guiding its use. A useful assessment would include the time spent by everyone who has to work with the result.

This is one way to understand the concern about AI content overload. Even accurate material can become burdensome when there is more of it than people have reason or capacity to absorb. The people receiving it also need room to think, respond and do things beyond consuming information.

What someone calls “AI-fatigue” may include several experiences. They may be tired of repetitive phrasing, long explanations with little substance, or uncertainty about whether the sender has considered what the words mean. A polished message can leave a reader wondering whether anyone will take responsibility for it.

Research on news offers one specific view of these attitudes. A 2025 Reuters Institute survey across six countries found that 43% of respondents were comfortable with news produced mostly by a human with AI assistance. The figure was 21% for news produced mostly by AI with human oversight. These statistics do not indicate how widespread AI-fatigue is, but reflect attitudes towards journalism. They suggest that the role people play in producing information matters to audiences. Reuters Institute, 2025.

Careful use of AI can also help people express something they mean. Someone writing in a second language might use it to find clearer wording. A person who finds writing difficult might use it to organise thoughts they have struggled to communicate. Judging sincerity solely by whether assistance was involved would overlook those possibilities. It matters whether the person understands the message, endorses it and can stand behind it.

There is evidence that AI assistance can save time on particular tasks. In a 2023 randomised experiment involving 453 college-educated professionals, access to ChatGPT reduced the average time spent on assigned professional writing tasks by 40%. This finding focuses on the tasks in the experiment. It does not establish that the participants gained more leisure or better wellbeing due to time saved. Noy and Zhang, 2023.

What happens to saved time is a separate decision. A workplace could use it to make workloads more manageable, allow deeper attention to difficult problems or shorten the working day. It could also raise the amount of work expected. Employers and workers still have to decide how the time saved will be used and who will benefit.

This matters because people have different amounts of control over their time. Someone facing insecure work or heavy caring responsibilities may have little freedom to turn a faster task into rest. Employers and institutions share responsibility for the conditions under which technological improvements become useful to the people doing the work.

AI can support meaningful living through the possibilities it opens as well. A learner could ask for an explanation at a pace they can follow, then try the problem themselves. Someone preparing for a difficult conversation could use it to organise their thoughts before speaking directly with the other person. In these examples, the value would lie in what the person becomes better able to understand or do.

Some effort remains worth protecting. Working through an argument can help us discover what we believe. Trying to explain something in our own words can reveal what we have yet to understand. When using AI, we can ask which part of that process we want help with and which part we want to practise ourselves.

A 2025 survey of 319 knowledge workers found that greater confidence in generative AI was associated with less reported critical thinking. Participants also described work shifting towards checking and integrating AI responses. Because the study relied on self-reports, it does not demonstrate that AI causes lasting loss of thinking skills. It does give us reason to examine how we use the assistance and what judgment the task still requires. Lee and colleagues, 2025.

We propose that learning to use AI responsibly should include recognising attention as a limited resource, both our own and other people’s. Before using a tool, we can name what we hope it will help someone understand or do. Afterwards, we can consider the effort involved for everyone affected, including our own time spent prompting and correcting. Learning, exploration and play can be worthwhile uses of that time. The aim is to recognise when the activity serves a purpose we value and when it is adding demands without helping us.

Schools, libraries and workplaces could offer practical exercises around this question. In a workshop, participants could prepare a short community announcement with and without AI, using the same supplied facts. Another participant could try to find the information they need in each version. Together, they could compare the time spent drafting, checking and reading, and discuss which version helped the reader understand what to do. This would be an exercise in noticing benefits and burdens, with room for AI assistance to be useful in different ways for different people.

Education could also include deciding when a task is finished. A person could identify what a useful result needs to contain before asking for suggestions, then pause to assess whether further revisions would add anything. For learning tasks, they could practise explaining the result in their own words or trying a similar problem themselves. These are practices we propose exploring, with feedback from the people using them.

Organisations can support that learning by agreeing how saved time will be used and asking staff whether AI has reduced their workload overall. Time spent checking outputs and answering other people’s questions belongs in that assessment. If the work has simply moved to someone else, the arrangement needs reconsidering. People receiving AI-assisted work should have a way to say when it is creating unnecessary demands on their attention.

Meaningful living will look different for each person and at different times in their life. It may involve learning a skill, caring for a relative, making something with friends or having an unhurried afternoon. People should have a say in what they want technology to make possible for them.

When AI helps someone finish a task, we should also ask what becomes possible in the rest of their day and whose effort that help requires. Time saved could become an unhurried conversation or the chance to put work down for the evening. The person whose life it is should have a say in what comes next.