The CHOICE System

The vision

Theodore Zeldin, in his book "An Intimate History of Humanity" (1994), wrote an inspiring chapter headed "What becomes possible when soul-mates meet". His vision of the quality of "humanity" is "an ideal of caring and kindness extending to every age and to every living being." Zeldin sees "humanity as a family that has hardly met", and he sees "the meeting of people, bodies, thoughts, emotions and actions as the start of most change." Describing such a meeting, he explains that the people he wrote about "were not soul-mates in the ordinary, romantic meaning of that word, but each owes the other the sense of direction which guides their life today."

At the very same time, in the couple of years leading up to the publication of Zeldin's book, the idea of the CHOICE System (CHOICES) was forming in my own mind. CHOICES offers a means for people to find, screen and contact other corresponding people who they may yet not know of. Correspondence means when both correspondents, at the same time, give suitable answers to each other's questions. And this is a means, I believe, of enabling that kind of meeting—of soul-mates in both senses—in many more people.

CHOICES promises to be much more effective than existing recruitment or introduction services, while being more secure by very nature. While current social networking services let people see public aspects of others who are connected, CHOICES instead lets people find correspondents, whether or not they are connected, based on what matters to them, and whether or not their concerns are out in public.

The number of questions that could be relevant to people making up their mind whether they want to meet is vast. CHOICES gives a way of creating and managing these question collectively and transparently, in a way completely different from recent developments in AI, so that people can ask and answer just those questions that either person finds most significant.

The most exciting potential lies, as Zeldin hints, beyond recruitment and dating. CHOICES provides a straightforward way of finding people with a common interest in anything at all, including such significant areas as sharing engaging visions, starting new ventures, or sharing living arrangements. It fills a gap in supporting an embryonic enterprise, before it has been properly set up. And it caters very well for individuals acting free lance and offering services to others for money — an area which is currently either fragmented, or when it is centralised on platforms like Uber or TaskRabbit, the central business extracts profit while having to do nothing beyond setting up the original platform.

Addressing practical challenges

By 1993, at City University London, the problems with an ordinary database approach to search had become clear to me. People querying a database of things, in an open marketplace, is fine, because the things do not care who is looking, and their owners don't mind who buys them. But it is quite different for people who are being searched for. What is needed, for the technology to help people find other people, is a system where the search happens both ways round at the same time—I am looking for people like you, at the very same time as you are looking for people like me.

Do people even know, at first, how to ask for what they really want, or even what it is they really want? Particularly when searching for other people, there may be a lot of uncertainty about exactly what kind of person you are looking for. To help with this, CHOICES gives simple, immediate and effective feedback about which of the existing questions it is most helpful to ask, and which to answer, to find people who correspond well.

No one wants to be reinventing the wheel, so it makes sense for CHOICES to provide a basic set of common questions to get people started. On the other hand, no one could possibly compile all the questions that may count as important for someone. A broad range and scope of questions is needed to address personal values and identity, often so important in finding suitable people. So, CHOICES allows users to create their own questions, which means that the range and scope of questions and answers is limited only by people's imagination.

There exist, out there, questionnaires that claim to test people's personal values. But there is no clearly agreed definition even about what personal values are, and certainly no standard questionnaire that covers everyone's values. So what questions would people actually use if they wanted to find others with compatible values? My belief is that even ordinary everyday questions, that people naturally ask each other in everyday conversations, are full of values implications, and treated intelligently, these everyday questions can be subtly effective in bringing together people who fit.

This gentle and subtle approach is especially important for people who feel their identity is marginal in mainstream society. People with uncommon values, preferences, or choices frequently end up hiding those values from public view, as they do not want to be discriminated against. The closer values are to someone's core identity, the more potential risk there is in disclosing them to others. With CHOICES, the only people who ever see sensitive information about someone are those whose answers correspond to the sensitive person's own questions. Thus, sensitive people can build in as many safeguards as they choose.

Trust is a vital factor in turning strangers into known and valued peers, and people may not initially tell the truth about themselves. To motivate trustworthiness, CHOICES keeps a track of all changed answers, while making honesty safe, by ensuring appropriate privacy. After people have been contacted, they are free to leave feedback about the credibility of the answers that were given to their questions. Where users are not being honest, this will quickly show up, and be viewable by other people, to warn them off.

The symmetry of the system embodies equality and fairness. Both sides can start out asking for what they really want. They may find someone corresponding immediately, or if not, they may choose to wait for someone who has what they want, and wants what they have. Alternatively, they can choose to make their wants more realistic, until they find someone who is already there. The same kinds of choice exist on both sides.

CHOICES respects people's time and effort, both asking and answering questions. Once a question is asked by someone, it is available for all to ask. CHOICES helps people to find and reuse questions as much as possible, and popular questions will rise to the top of the list. Specialist questions are still there, but they will only be asked if it is calculate that they are relevant. CHOICES prompts you to answer questions that are most relevant to finding correspondents. Once you answer a question, you can reuse the answer in any other enquiry without having to answer again.

Two distinct processes will help manage the rich diversity expected from letting people ask their own questions.

  1. Interested people can propose improvements to the wording and the structure of user-defined questions. These will be put to other interested parties, and where it seems that there is general agreement on a proposed improvement, it will be adopted.
  2. The system itself will monitor the performance of questions, measuring the information generated. Questions that generate most information will rise to the top of the list of questions to ask and answer. This will use entropy measures in ways related to machine learning.

A very large number of people will have access to CHOICES through their smartphones as well as computers. They will be able to find others to help each other to renew their hopes and lives, with very little risk of being found by the "wrong" people. As Zeldin figured, this is a great way to enable change, in a society that dearly needs positive change.

How might this work in practice?

As currently envisaged, the system groups enquiries into six major areas:

The first and the last of these tap into the already highly lucrative market for jobs and personal ads.

People make enquiries under one of these headings, and they add questions to each enquiry, choosing the answers they will accept from the other person. The system prompts them with the relevant questions asked by other people that they need to answer in order to find correspondence. If they don't find who they are looking for straight away, their enquiry is left open for others searching under the same heading to find correspondence with. When there is correspondence on both sides, people are offered the option of seeing each other's contact details, but only by mutual consent. For a beginner, it can be a very simple process of answering the relatively few relevant questions that are suggested; while a more confident user might dig deeper to find and ask more probing questions, or even to compose a question if the existing ones don't offer sufficient discrimination.

The original name "CHOICES" was devised as an acronym standing for "Common Human Online Information Correspondence Enquiry System". Perhaps this is as accurate and relevant now as it was right at the beginning.

More information