RX-4 Recalibration Review

RX-4 was built to detect survivors, assess casualties, preserve telemetry, and support extraction teams in environments where delay can become fatal. Years of retained operational memory have made the droid more capable at those tasks—and less standardized in how it performs them.

RX-4 increasingly rechecks uncertain life signs after approved casualty models favor closure, preserves contradictory telemetry rather than simplifying it, and compares present conditions with earlier recoveries whose outcomes were not known when the original decisions were made. Some of those judgments have helped identify survivors. The same behavior can prolong hazardous operations, delay other decisions, consume limited resources, and place unusual weight on the capabilities of Brina Tenebrix.

Czerka Interstellar has therefore opened a formal recalibration review. The question is not simply whether RX-4 works. It is whether the accumulated experience responsible for its unusual value can be separated from the divergence now causing concern—and what must survive before any irreversible change is considered.

The RX-4 Recalibration Review exists to determine whether the droid's increasing divergence from approved casualty-assessment models represents useful learned judgment, unacceptable operational bias, correctable technical drift, or some combination of those conditions.

A full reset is not a predetermined result. Continued operation, narrower modification, additional monitoring, restricted functions, reassignment, and more extensive recalibration remain possible depending upon what the evidence establishes.

The review must answer two related but separate questions: what changes, if any, are necessary for RX-4 to remain a defensible recovery asset, and what parts of its accumulated memory, telemetry, and experiential continuity must be preserved before an irreversible intervention occurs.

• Comparative casualty accuracy — When RX-4 and approved casualty models disagree, which assessment is better supported by eventual independently established outcomes?

• Verification cost — Do additional scans and reassessments improve outcomes enough to justify longer hazard exposure, delayed closure, or diverted recovery resources?

• Brina-specific weighting — Does RX-4 accurately account for a demonstrably capable recovery partner, or has repeated partnership distorted recommendations that should generalize beyond Brina?

• Historical inference — Does RX-4 reliably distinguish evidence observable in the present case from patterns inferred through earlier recoveries?

• Sensor retrofit — Does the Czerka-adapted Zakuulan-origin multispectrum sensor modification installed in 3622 BBY materially contribute to the present variance, and how much can incomplete provenance establish?

• Memory preservation — What telemetry, configuration data, learned associations, and operational history can be independently copied and authenticated before modification?

• Investigative relevance — Does RX-4 retain evidence material to Meridian Recovery, the Tenebrix disappearances, or Grey Ledger correlations that must remain reviewable under defensible custody?

• Alternatives to destructive recalibration — Can demonstrated risks be addressed through monitoring, restricted functions, narrower modification, separated decision models, reassignment, or another reversible measure?

RX-4's memory creates a preservation problem that ordinary data copying does not automatically resolve.

Raw telemetry might be copied without preserving every association through which the droid compares one recovery with another. Software and configuration could potentially be preserved while learned weighting, relationships, behavioral adaptation, or the experiential patterns connecting years of observations are altered or lost.

The established record does not identify a verified backup known to restore the complete experiential continuity that would be lost through a destructive reset. That does not establish that such restoration is impossible. It means the review cannot assume that retaining data and restoring RX-4 would necessarily be equivalent.

The same distinction allows different characters to approach the review without sharing one philosophy of droid personhood. Czerka may treat RX-4 as corporate equipment. Brina may understand its retained continuity as belonging to an individual she has worked beside for years. Bella can argue against irreversible destruction because active evidence must remain reviewable without resolving the personhood question at all.

Not every disputed behavior can be meaningfully tested inside a maintenance bay.

Controlled diagnostics can measure sensor response, model variance, software behavior, and technical integrity. They cannot completely reproduce a collapsing structure, ambiguous life signs, incomplete medical information, a shrinking extraction window, or the capabilities of the operative actually attempting the recovery.

Legitimate field assignments therefore provide important review evidence when RX-4 and standard closure criteria produce materially different recommendations and the eventual outcome can be independently examined.

Those assignments should arise from genuine operational need rather than danger created merely to test the droid. RX-4 may identify a survivor others would have abandoned. It may prolong a search and find nothing. Brina may complete an extraction that appeared statistically unreasonable. Another team may show that the same recommendation depended upon assumptions specific to Brina.

The RX-4 Recalibration Review is restricted Czerka continuity. The complete reasons for the review and the evidence under examination are not public knowledge.

The review succeeds when it produces a decision the surviving evidence can defend. That decision does not need to preserve RX-4 unchanged, and it does not need to result in recalibration.

Possible outcomes should follow evidence and roleplay rather than assume in advance that preservation, modification, or reset is the morally correct result.

• Continued operation with accumulated memory intact and additional monitoring.

• Limited recalibration addressing a demonstrated problem while preserving evidence and experiential continuity as far as technically possible.

• Separation of Brina-specific recovery modeling from general casualty recommendations.

• Restriction or reassignment of particular RX-4 functions while other recovery duties continue.

• Independent preservation of telemetry or memory material before further modification.

• Suspension of destructive recalibration while material Meridian or Grey Ledger evidence remains inadequately preserved.

• More extensive recalibration after narrower interventions prove insufficient and the evidentiary consequences have been addressed.

• Continued physical preservation of RX-4 with reduced or suspended field deployment.

• A technically defensible result that leaves droid personhood unresolved.

Failure can also take more than one form. Irreversible intervention can destroy evidence before anyone understands what was lost. Attachment can substitute for analysis. Liability management can substitute for technical judgment. Reassessment can continue until uncertainty itself exposes other people to unacceptable danger.

For RX-4's complete operational history, capabilities, relationships, and current circumstances, see RX-4. RX-4, Lorim Vance, Bella Tenebrix, and Brina Tenebrix provide the corresponding character records for the review's principal human participants.

The wider recovery, missing-person, corporate-custody, and aftermath story continues through Meridian Recovery. Grey Ledger covers the broader cross-institutional pattern of contradictory records with which some of RX-4's retained telemetry intersects.