AI for Science
Client-Controlled Analysis: Keep Source Data Under Owner Custody
Published 2026-08-22 · Updated 2026-08-22
Answer in brief
The practical question behind Client-Controlled Analysis is which workload and output can be approved without transferring the source dataset. Rank credible alternatives with data location, workload package, key release, attestation, logging, output review, and deletion, expose the strongest counterargument, and challenge the leader by trying to prove the analyst cannot access raw inputs through tools, logs, or outputs. A useful answer changes the next allocation decision without pretending computation is final proof.
Evidence status: Decision-method guide; not a completed investigation or final validation.
The decision this guide supports
which workload and output can be approved without transferring the source dataset
Why the problem is difficult
The article-specific identification challenge is whether the question “which workload and output can be approved without transferring the source dataset” can be resolved using data location, workload package, key release, attestation, logging, output review, and deletion, rather than merely restated in new language.
A falsifier-first workflow
- Define the decision precisely: which workload and output can be approved without transferring the source dataset.
- Build a source and data ledger around data location, workload package, key release, attestation, logging, output review, and deletion.
- Compare the inherited route with a mechanistically distinct alternative and a constraint-based null.
- Actively search for the strongest counterevidence relevant to this decision, including boundary cases and prior failures.
- Run the lowest-cost discriminating challenge: prove the analyst cannot access raw inputs through tools, logs, or outputs.
- Record pursue, reframe, or stop, the confidence level, the evidence ceiling, and who owns downstream validation.
Decision criteria
- Decision impact: would the result materially change the choice about which workload and output can be approved without transferring the source dataset?
- Evidence fit: does the available evidence—data location, workload package, key release, attestation, logging, output review, and deletion—directly address the decision rather than merely correlate with it?
- Discrimination: does the preferred route predict an outcome a credible alternative does not?
- Robustness: does the ranking survive the challenge “prove the analyst cannot access raw inputs through tools, logs, or outputs”?
- Validation boundary: is the conclusion no stronger than the available sources, data and computation?
Supporting evidence
data location, workload package, key release, attestation, logging, output review, and deletion
Counterevidence
For this decision, a result from “prove the analyst cannot access raw inputs through tools, logs, or outputs” that reverses or flattens the ranking must remain visible even when it is commercially inconvenient.
Computation
Here computation earns its place only if it changes the choice about which workload and output can be approved without transferring the source dataset or exposes why the available evidence cannot resolve it.
Fastest falsifier
prove the analyst cannot access raw inputs through tools, logs, or outputs
When to stop or reframe
A decision-specific stop trigger is failure of the challenge “prove the analyst cannot access raw inputs through tools, logs, or outputs” without an independently supported alternative mechanism.
Evidence ceiling
Owner custody still requires secure implementation, lawful processing, and competent oversight.
Sources and starting points
- NIST AI Risk Management Framework — A voluntary framework for governing AI risk, measurement, and accountability.
- Google Cloud: Confidential Space overview — A documented separated-role architecture for attested confidential workloads.
- C2PA specifications — Open technical specifications for content provenance and authenticity metadata.
- OWASP: LLM Prompt Injection Prevention — Defensive guidance for treating retrieved and user-provided material as untrusted input.
- OWASP Machine Learning Security Top 10 — Additional authoritative starting point selected for this decision area; applicability must be checked against the precise question.
- NIST Generative AI Profile — Additional authoritative starting point selected for this decision area; applicability must be checked against the precise question.
Continue the decision journey
- AI for Scientific Discovery: Where It Helps and Where It Fails
- Multi-Model Consensus Is Not Independent Scientific Replication
- A Frontier-Model Research Workflow With Human Accountability
- The Evidence Ceiling in AI-Assisted Science
Explore the full topic hub · Editorial standard · Scientific Oracle consulting
Frequently asked questions
- What decision does “Client-Controlled Analysis: Keep Source Data Under Owner Custody” help make?
- It supports a bounded decision about which workload and output can be approved without transferring the source dataset. The framework keeps alternatives, evidence, counterevidence, uncertainty, and the fastest falsification test visible.
- What is the fastest useful test?
- prove the analyst cannot access raw inputs through tools, logs, or outputs
- Can computation validate the final scientific claim?
- No. Owner custody still requires secure implementation, lawful processing, and competent oversight. Computation can prioritize and eliminate directions; final validation remains with the appropriate domain methods and accountable specialists.
- When should the project stop or reframe?
- Stop or escalate to human review when sources cannot be verified, prompts or retrieved files may be malicious, output changes materially across reasonable runs, protected data handling is unresolved, or the model is being asked to make a regulated or final scientific decision.