Augmented Learning Collaboration (ALC) - a new paradigm for digital togetherness

Augmented Learning Collaboration (ALC) will enable better individual collaboration habits and collective collaboration practices in all domains of life. By better we mean more intelligent, effective, efficient, trustworthy, motivating, and inclusive. The aim is thus to progress individual quality of life and common solidarity in society. But what do we understand by ALC? A primer on some of the key concepts such as symFlos, symBots, virtual private assistants...

SymFlos, symBots, VPAs (user, experience) and Extended Identity

At SymPlace we see any collaborative undertaking as a symFlo, i.e., a flow combining different types of structured interactions, to guide and / or drive the collaboration. SymFlos are configured to consider key issues such as human cognitive specificities, social interaction, complementary competencies, time factors or data ownership. Our long-term vision is that in the future, instead of emailing, messaging or writing on cards, we will “symly”, i.e., invite each other to such symFlos and then interact through them. And that instead of attending meetings we get invited and participate in symFlos where the meeting becomes an event that involves those people that are motivated and competent to contribute to the open issues that need discussion. And that instead of ending up with long list of messages, cards and / or workspaces, we are in control of the symFlos which we have chosen to initiate / lead or to participate in or to observe. Participants will be invited and agree to participate in a symFlo; depending on their role they will receive more or less information. We will symly instead of having to organise and take care of multiple (at best) loosely coupled threads of messages and actions.

Many such structured interactions will be complex aggregates composed of “basic” generic symFlos: e.g., issuing and accepting invitations or proposals; raising and answering questions; voting, choosing and deciding; specifying who is involved and how (role in the collaboration; responsibility with respect to the outcome…; anonymously or nominatively; in real-time or asynchronously; fast paced or in slow motion…; informal exchanges (chat, videoconferencing…). Collaboration is not necessarily only digital and structured: symFlos can include small face-to-face meetings or videoconferences, with dedicated steps for registration and collecting notes. Transactions and payments can help to contractualise commitments, reciprocate efforts and /or exchange money. SymFlo templates represent reusable patterns of such composite symFlos, corresponding to collaboration practices and / or habits.

Elaborate symFlos create more administrative overhead to drive the process and to orchestrate the involvement of participants. SymPlace adds symBots to the game, i.e., virtual agents which take care of the nitty-gritties; they guide the participants, invite them, remind them to contribute, consolidate information, archive results, help users to act when needed. Each symBot embodies a social practice for collaboration, and with it the logic for coordinating different intelligences, the collective agreements to follow an established practice and the trace of agreements and other commercial transactions included in the symFlo. SymBots can also encapsulate and hence integrate third party technologies such as Large Language Models (such as ChatGPT4), neural deep leaning algorithms and data sciences. SymFlo templates map can be considered as the genotypes of symBots, and symBots as the assistants to support the participation in the symFlo instantiated from the symFlo templates.

Each user will be assisted by a Virtual Private Assistant (VPA), a special symBot in charge of helping the user to manage his.her participation in all the different symFlos and actions s.he is involved in, together with the corresponding data. The VPA also orchestrates the cooperation with other virtual agents, and in particular other users and their virtual private assistants, as well as the user’s commitments and future actions. One of the core ideas behind the ALC approach is that symFlos will require a new form of user experience, centred on digitally assisting them in coping with multiple, complex symFlos, in multiple domains, over extended periods of time. At SymPlace, first experiences with a first product, followed by prototype developments including user testing, have highlighted the UX challenge of ALC. Therefore, every user will have his dedicated “idiosyncratic” VPA. By proposing different views, with more or less simple versions, the VPA guides its “human master” through symFlos, helps her.him to stay on top of all his symFlos, organises her.his different contributions (e.g., responding to invitations, providing his inputs, participating in decision/making processes etc.), manages the selfSymFlo, i.e., the memory / log of her.his activities in the different symFlos and the respective identities s.he may have. This approach makes it possible to organise the information the user sees and the modes of interactions, according to her.his preferences, experience (with the app and with specific symFlos) and role. Universality is a leading principle of the SymPlace vision: one Virtual Private Assistant per user helps to interact with the virtual agents of the entire ecosystem. This approach also simplifies the challenge of extending the user interface to speech recognition (it is much more logical and effective to position speech recognition within the VPA and thus, combined with natural language processing, endow it with a much more intimate knowledge of the speaker-user).

Thus, in summary, collaboration between humans is augmented by symBots (including the VPAs).

All symBots will have selfSymFlos, i.e., a lasting “memory” which stores the transactions (blocks connecting, in a tamper safe mode, the necessary information connecting different selfSymFlos of cooperating VPAs / symBots), which corresponds to contractual agreements between the collaboration parties (including possibly trusted third parties when there are commercial transactions). The duration of these memories is determined by the rules governing the symBot (and can range from the time of a specific collaboration to unlimited). The memory of symBots will condition how long the respective information will be available and exploitable. The selfSymFlo of the VPA of all transactions and interactions by its master becomes the Extended Identity of this user. This Extended Identity can then be used in cooperations between the VPA and neutral trusted third party symBots, as a mean to verify important facts about a person to confirm her.his trustworthiness, without ever sharing and hence compromising her.his personal data, or even her.his identity. It is also the basis for learning from past experiences.

The “DNA” of symBots will correspond to the practices that they support. Practices which can be evaluated, combined, improved, reviewed, certified… by using metaSymFlos. Thus, will emerge an ecosystem of collaboration practices, which will provide an increasing wealth of solutions for better collaboration at work, in education, in health, in democracy, in leisure, in personal life… And symFlos can become a powerful mean for organisations and communities to establish common practices, implement, govern and continuously improve them.

I.e., we create a system which supports learning at the level of individual users and individual symBots, but also of the collaborating system seen as a whole. And that is the essence of Augmented Learning Collaboration.

Some symBots will embody the interaction with third-party information tools and systems. This includes symBots encapsulating different services and algorithms.

The first application will be specialised LLM to provide multiple services, such as

  • Expert symBots contributing vast domain expertise, from selected perspectives and other symBots helping to check the quality of the input from these “virtual experts”.
  • MetaSymBots, i.e., symBots for metaSymFlos, to scout for interesting symFlo templates for a given collaborative undertaking and help to users to choose the most relevant one.
  • MetaSymBots to evaluate symFlos and suggestion improvements to the corresponding symFlo Templates.
  • MetaSymBots generating automatically new composite symFlo templates, from the interaction with symFlo designers, possibly from natural language requests.
  • ...

Other application include:

  • Recruitment symBots to identify and recruit human experts who are particularly proficient / successful in the usage of a symBot.
  • SymBots to automatically associate Tags to contents collected by other symBots, to prepare information to be used for further action (e.g., to produce an action list or minutes, which can then be used by other symBots to obtain agreement on the summary).
  • Access to data sciences to aggregate and exploit data provided voluntarily for specific purposes, and providing feedback to individual contributors, including those aggregated data and the meaning for them. Such symBots would for instance be used to correlate user specific quantitative data with big data, to provide useful individual feedback and guidance, e.g., for applications in domains such as health (e.g., fine-tuning therapies), educational (e.g., guiding students with special learning abilities / difficulties) or finance (e.g., evaluating risks and opportunities with special investments).
  • System and application programming interfaces to information systems and dedicated applications such as customer relationship management, human resources management, content management systems etc.
  • ...

Use case: the (un)happy hyperCol

Let’s consider the splendours and miseries of Jack, a prototypical business consultant. He is what we call a “hyperCol”, i.e., somebody with multiple participations in collaborative undertakings. He spends a lot of time in the “hyperactive hive”: reading and responding to incoming messages, following up on unanswered messages, interacting with people in meetings and on phone calls to discuss things which otherwise could be processed much more effectively by asynchronous means … leaving less time for productive work. A lot of this work is just noise processing (sorting the more from the less important stuff) and activity prioritisation (organising information to process them effectively). The more he is chased by messages following other messages, the higher the probability that he we will not respond. The cognitive overload generated by these activities switching back and forth and the resulting stress are a problem in many work environments, and especially for such consultants and designers such as Jack.

The symlying workday of Jack would be quite different:

  • In the morning Jack checks by domain of activity the status of symFlos he leads. Jack focuses first on those requiring his intervention as superCol. As needed, he invites other participants to new symFlos (using corporate symFlo templates) and / or adjusts existing ones, possibly adding new symFlo templates.
  • He then contributes to those where he is committed. For the more unusual symFlos he lets the respective symBots guide him. For the usual ones, he uses the high-speed version which allows him to get things done in record time.
  • He checks on new invitations which he accepts / declines / postpones to later (when his VPA will remind him). Some of these are generated by this company’s information management system, which has pushed some customer requests to him.
  • He holds a few meetings as part of ongoing symFlos.
  • He checks on highlights of symFlos which he “observes” and receives from one of the team symBots a summary of outcomes of a meeting, which he did not have to attend.

The symlying Jack is more reliable, less stressed, more productive, more successful in his undertakings, better liked by everybody…