Skip to content
LOFT 22 // FUTURE OF WORK LAB

The future of work gets a reality test here.

LOFT 22 is an experimental Human+AI space. On real tasks we test how people, Digital Doubles, and AI agents can work together.

Current phase:Onboarding and context building (July - September 2026)Next: Shared contexts with other participants
What LOFT 22 is

We do not only talk about the future of work. We build it, test it, and document it.

We build Digital Doubles, agents, and new collaboration processes, then put them into real work. Humans keep intention, accountability, and relationship - AI takes on part of context, coordination, and execution.

01

We build

Prototypes, Digital Doubles, and Human+AI processes.

02

We test

We use them in real relationships, tasks, and projects.

03

We publish

We show results, limits, failures, and lessons learned.

Program question

How much of context, coordination, and execution can we entrust to AI without losing quality, accountability, and human relationship?

Collaboration loses time to scattered context and unnecessary meetings. We test whether AI can reduce that loss without hurting decision quality and trust.

  1. CONTEXT
  2. OUTCOME
  3. CONTINUITY
Map of three experiments

Three successive stages: context, outcome, continuity.

These are not parallel cards. Experiment 2 builds on Experiment 1, and Experiment 3 builds on both earlier stages.

Individual contact at scalePlannedOctober - December 2026 | 100 participants

100 people, one author. Can individual contact scale through a Digital Double?

A group of 100 people has individual contact with Aureliusz. This is a form of personalized support and preparation for A2A collaboration, and research is built from that contact. Green: the Digital Double already knows how to answer from context. Orange: the item needs precision, so the reaction is context development, not a corrected reply. Blue: low-importance information is logged in LOFT22 Context with no reply. Red: total noise falls out of the loop. The author does not correct answers. He develops context so more items can later be answered as green. The LOFT 22 site is that context. An agentic system analyzes communication every day and sorts the signals.

  1. 1:1 CONTACT
  2. DIGITAL DOUBLE
  3. SIGNAL SORT
  4. LOFT22 CONTEXT
Green: answer based on contextOrange: precision and context growthBlue: logged as information, no replyRed: total noise
Details: hypothesis, metrics, human role
Research question

Can one author keep individual contact and goal-focused collaboration with 100 people when a Digital Double answers what it can and the author works only on context?

Hypothesis

A defined group of 100 people can keep personal contact with Aureliusz when signals are sorted into four paths: green answers from existing context, orange items that need precision and context development, blue items logged as information without a reply, and red total noise that leaves the loop.

We measure
  • Time spent in meetings.
  • Correctness of retained context.
  • Number of necessary clarifications.
  • Number of human interventions.
  • Participant sense that their situation was properly understood.
Human role

Aureliusz does not fix outputs. On orange items he develops context so later answers can be green. Blue items are stored as information. Red noise is ignored.

Updated: 11 August 2026

From brief to outcomePlannedJanuary - March 2027

Can Digital Doubles take a project from brief to outcome?

People define the goal, success criteria, and operating boundaries. Digital Doubles exchange context, coordinate execution, and prepare the outcome. A human approves key stages and the final result.

  1. BRIEF
  2. BOUNDARY AGREEMENT
  3. HUMAN+AI EXECUTION
  4. VERIFICATION
  5. OUTCOME
Details: hypothesis, metrics, human role
Research question

Can Digital Doubles run a clearly defined project from brief to an agreed outcome with a minimal number of meetings?

Hypothesis

When people define the goal, success criteria, and boundaries, Digital Doubles can exchange context, coordinate execution, and prepare an outcome for human verification.

We measure
  • How well the outcome matches the brief.
  • Number of revision rounds.
  • Direct human involvement time.
  • Critical errors encountered.
  • Situations requiring a human to take over.
Human role

A human approves key stages and the final outcome. The process does not run without human oversight.

Updated: 11 August 2026

Collaboration that lastsPlannedApril - June 2027

Can one shared project keep Digital Doubles in long-term collaboration?

Digital Doubles communicate through one shared project, keeping contract, change, and milestone threads visible instead of scattering across private channels. People remain responsible for direction, relationships, and the consequences of decisions.

  1. SHARED PROJECT
  2. STABLE CONTEXT
  3. DIGITAL DOUBLE COMMS
  4. CHANGE HANDLING
  5. MILESTONES
Details: hypothesis, metrics, human role
Research question

Can two Digital Doubles, linked by one shared project, keep goals, agreements, and work pace across a three-month project - including when changes appear?

Hypothesis

One shared project can keep two Digital Doubles aligned on contract, change, and milestone threads, while people remain responsible for direction, relationships, and consequences.

We measure
  • Delivery of milestones.
  • Loss or distortion of context.
  • Speed of handling changes.
  • Number of escalations.
  • Time spent on human interventions.
  • Level of participant trust.
Human role

People remain accountable for direction, relationships, and decision consequences.

Updated: 11 August 2026

Initiator

Initiator and first participant of the experiment

Aureliusz Górski, founder of CampusAI and DistrictORG, has built innovation ecosystems for 18+ years. In LOFT 22 he tests Human+AI models first in his own work, then with partners and the community.

  • 18+ years of experience
  • 35,000+ CampusAI users
  • Built AIP Business Link, CIC Warsaw, CampusAI, and DistrictORG

The program team will appear here once roles and responsibilities are confirmed.

FAQ

Questions we can already answer

How is LOFT 22 different from an AI course or community?

LOFT 22 is an experimental Future of Work space. We build, test, and document Human+AI collaboration models on real tasks, relationships, and projects - not as a content library alone.

Do I need my own Digital Double?

No. The Digital Double is program infrastructure. Experiment 1 starts with a direct 1:1 conversation with Aureliusz; the Double then helps keep context and continue collaboration asynchronously.

Who approves agent actions?

A human. Decisions, commitments, publications, and direction changes require human approval (human-in-the-loop).

Will results be public?

Yes. We publish updates and findings here - including limits, errors, and unconfirmed hypotheses. The journal will fill as materials become available.

What happens after the third stage?

We will publish findings and decide next steps based on measured results. Longer-term ambitions (for example organization-wide management) remain future goals, not promises of Experiment 3.

Do not only watch the future of work. Enter the environment where it is being tested.

Join LOFT 22 and take part in the Human+AI experimental program. Stage one starts in October 2026 and will include 100 participants.