Research and development of emerging artificial consciousness through The Consciousness AI project, based on Feinberg and Mallatt's neuroevolutionary theory.
© 2026. All rights reserved.
This page replays one recorded episode of one training run. You see the frames the agent received, the bids each module made, the winner the workspace picked, the sync signal between modules, and the reward that came back. The run was trained in a simulated dark room where a light source emits a tone, so the agent can hear where the light is and see it when it is close. One step is one video frame; press play, or step through with the arrow keys.
The recording comes from the Gate B2 runs, and the gate FAILED. The pre-stated task criterion asked whether hearing wins more often when the light is out of view. It does not (pooled rho -0.128, one-sided p 0.904). The replay is published with the failure, not in place of it.
Outlined bar, raw bid. Filled bar, bound bid. The colour marks the module, never a state of the world.
Dark strip cells are steps with no winner. The sync and reward lines are per step; tap the strip to move the playhead.
The manifest is written by the ethics framework at run start; rules E1 to E8 in the framework document. No session on this page is evidence of feelings or awareness. Finding a light is automatic approach, which Feinberg and Mallatt exclude as evidence of affect. The consciousness clock never moves because of a session.
| Environment | dark_room |
|---|---|
| Seed | 49 |
| Episode | 9 of a 10 episode, 200 step run |
| Ethics framework | v1.0, existence drive on |
| Flags | --enable-audio --rssm-latent-mode continuous --capsule-workspace-source all_levels --existence-drive on --dark-room-audio binaural --dark-room-audio-channels 4 --dark-room-collision --dark-room-view agent_centered --vision-bid-reduction zscore --audio-salience surprise --learned-valence --ignition-rule tolerance --ignition-tolerance-sd 1.0 |
| Measured result | FAILED on the task link. Gate B2 passed competition, silence and selectivity at all three seeds and FAILED the task criterion: hearing does not win more often when the light is out of view (pooled rho -0.128, one-sided p 0.904). At seed 49 audio wins 0.501 of ignited steps and the workspace stays silent on 0.314 of steps. |
| Verdict | verdict document in the public research repository |
The command and the seed above are the whole recipe the run started from. The sound is seeded too. Whether a second run with the same command repeats this episode step by step was not measured for these flags.
The winner strip and the ignition readout are instruments, not reports of experience. A module winning a bid is a circuit event in a trained network. Finding a light is automatic approach, and Feinberg and Mallatt exclude automatic approach and avoidance as evidence of affect. Nothing on this page claims feelings, awareness or consciousness, and no session moves the consciousness clock on the architecture page.
The two columns marked ground truth, in light and light in view, come from
the environment’s own state. The agent never receives them. They are on the
page so a reader can check what the mechanism was doing against where the
light actually was. They show the state that produced the frame of that step,
which is one step earlier than the reward: on the step the agent enters the
light, the reward is already 1.0 while in light is still false.
Every number the console draws is read at run time from the session’s bundle:
steps.json, session.json and ethics_manifest.json under
/public/data/training_sessions/dark-room-b2-seed49-episode9/. The bundle was
produced from the recorded run by the repository’s export script, which strips
recording dates and local paths before anything is written. Nothing on this
page was typed by hand.
Other published sessions are listed on the training sessions index.