The core proposition
From possibilities to a choice.
When available information and alternatives exceed a person’s perceived ability to compare them against their own priorities, the search for human mediation to reach a decision may increase. AI may foster this abundance directly or through proposals supplied by other people.
The starting point is a recognisable situation: many possibilities are available, yet identifying which matter personally feels difficult. Content quantity, number of alternatives and comparison difficulty are different concepts that research must distinguish.
Human mediation means a conversation that helps make criteria explicit, discuss trade-offs and recognise consequences. The decision remains with the person who will live with it.
One possible pathway, to be tested
What might happen?
- 01
Possibilities increase
Information, answers and proposals arrive from multiple sources. AI may contribute to their production through other people as well.
- 02
Comparison becomes difficult
Options may be plausible and attractive while personal priorities and trade-offs remain unclear.
- 03
I seek a person to discuss the choice
A conversation may help weigh criteria and reach a decision. Its contribution must be observed rather than assumed.
The pathway can stop or take another direction. AI may summarise, narrow the options and support independent choice. A person may confuse, pressure or merely confirm what we want to hear. These outcomes also define the hypothesis’s boundaries.
One example, many settings
Thirty renderings. One home to choose.
Imagine three architects each presenting ten renderings, assisted by AI. The client faces thirty images and may ask AI to compare them, receiving further variants, a ranking or a summary.
That may be sufficient. Alternatively, the client may still wonder: “Which design fits the way we live? What would I have to give up?”. Discussion with a professional may clarify space use, budget, maintenance and trade-offs. The reasons should be understandable and open to challenge.
This is an illustrative example, not a study finding. Image count alone does not establish overload; trust in a professional does not guarantee a sound proposal.
Personal decisions
Options for a purchase, a course or a project may call for clarification of the criteria that matter.
At work
Abundant proposals and presentations may prompt discussion of feasibility, priorities and responsibility with a colleague.
With a professional
A well-presented proposal still calls for questions, explanations and understanding of trade-offs.
With people close to us
A conversation may connect alternatives to personal circumstances while preserving disagreement and autonomy.
Explore the scientific foundations
Existing foundations. A proposition to test.
These references help distinguish concepts. The link between AI-enabled abundance and the search for human mediation remains the hypothesis under investigation.
Choice overload
Chernev and colleagues’ meta-analysis considers option complexity, task difficulty, preference uncertainty and the decision goal. It helps specify when a large assortment may impede choice. [1]
Many options do not necessarily produce difficulty or a search for another person.
Cognitive load
Cognitive load theory examines task demands, knowledge and learning. It provides context for thinking about processing difficulties. [2]
Reported effort in this survey is not a direct measure of cognitive capacity.
External aids and cognitive offloading
Risko and Gilbert describe using actions and external aids to reduce a task’s cognitive demands. This framework also encourages considering how a tool may ease comparison. [3]
The framework does not establish a general preference for human or digital assistance.
Trust and epistemic vigilance
Sperber and colleagues propose a framework for mechanisms involved in evaluating information communicated by others. It is relevant to assessing reasons and source reliability. [4]
An interlocutor may be persuasive and wrong: trust and accuracy must be distinguished.
Experience, knowledge organisation and neuroplasticity
Chunking and long-term working memory help describe the role of domain-specific experience in organising and retrieving information. Neuroplasticity studies show that experience can be associated with brain changes. [5] [6] [7]
These references provide cognitive background. The current study does not measure brain changes, neuroplasticity training or general improvements in learning. The new hypothesis does not depend on such outcomes.
Priorities · Dialogue · Responsibility
“What would you advise me to do?”
This question may seek competence, another perspective or help weighing alternatives. Useful dialogue allows questions about reasons, clarification of priorities and the freedom to disagree.
Voice, gaze, posture and silence can form part of a meeting. They require contextual interpretation and do not certify sincerity, competence or truth. The brief questionnaire does not measure their effects.
“We will embrace again” expresses a possible return to consultation during a present decision. The “thank you” afterwards conveys its human significance; research must distinguish feeling helped, reaching a choice and making an objectively better decision.
Testable conditions and explicit limitations
What can we observe today?
The brief questionnaire connects a real episode to distinct measures. Earlier submissions continue to contribute wherever items are compatible.
| Aspect | What we collect | Limitation to retain |
|---|---|---|
| Quantity and effort | Contents consulted and perceived comparison difficulty. | Contents may have different sources; feasible alternatives are not all counted. |
| Need and behaviour | Welcomed or necessary human input, actual pathway and direct consultation order. | Reporting a need does not establish that it was inevitable or caused by AI. |
| Direct and indirect AI | Personal AI use and declared or known use in others’ proposals. | Unknown provenance remains unknown; it is not inferred from appearance. |
| Contribution to choosing | Decision status and perceived change after human consultation. | Retrospective assessment does not establish a causal effect or decision accuracy. |
What would strengthen or weaken the hypothesis?
A consistent within-episode association between quantity, difficulty and seeking consultation would be pertinent. The perceived contribution of conversation provides distinct information. Separate marginal percentages do not demonstrate the whole pathway.
The hypothesis needs boundaries if quantity is not associated with need, AI alone clarifies priorities or human consultation does not improve perceived ability to choose. Greater quantity may itself reflect a choice that was already difficult: the survey cannot settle that causal direction.
Attributing an effect to information quantity requires a separate causal design to be considered after this phase; none is activated now. New items and twenty-submission reporting intervals do not make the analyses an independent confirmation.
Research is valuable when it requires revising the proposal. No answer is excluded because it contradicts it.
A conversation may help me understand which decision I am prepared to take responsibility for.
The revised core, on one page
From abundance to a decision
A visual summary of the proposed pathway, alternative pathways and research limitations.
Updated 31 October 2026. It represents the hypothesis, not a chart of study findings. Original infographic in Italian.
Selected bibliography
Studies and theoretical frameworks
A focused narrative overview, not a systematic review. These sources address different concepts and settings and do not automatically validate the overall proposition.
- Chernev, Böckenholt & Goodman (2015). Choice overload: A conceptual review and meta-analysis. ↗
- Sweller, van Merriënboer & Paas (2019). Cognitive Architecture and Instructional Design: 20 Years Later. ↗
- Risko & Gilbert (2016). Cognitive Offloading. ↗
- Sperber et al. (2010). Epistemic Vigilance. ↗
- Chase & Simon (1973). Perception in chess. ↗
- Ericsson & Kintsch (1995). Long-term working memory. ↗
- Draganski et al. (2004). Neuroplasticity: changes in grey matter induced by training. ↗
Scenes and cover include AI-generated images; they do not depict experiments or research testimonials. The infographic is a conceptual model.