Model & Inference

Implement confidence scoring to communicate output certainty alongside each result. Calibrate before deployment.

Define model accuracy acceptance criteria aligned to business requirements before validation commences.

Implement counterfactual explanation to show users what changes would alter the model's output.

Communicate model accuracy, known limitations, and uncertainty to users in the production interface at launch.

Monitor production accuracy continuously against the validated baseline. Trigger model review when accuracy degrades.

Specify a RAG architecture at design stage for factual domains. Define grounding requirements and acceptable hallucination thresholds.

Evaluate foundation model candidates on hallucination benchmarks at design stage. Select models with lowest documented rates.

Design system prompts to instruct the model to acknowledge uncertainty, cite sources, and refuse when knowledge is insufficient.

Fine-tune on a curated, domain-specific dataset to improve factual accuracy. Validate hallucination rates pre/post fine-tuning.

Configure conversation controls at deployment to restrict the model to approved topic domains and escalate off-topic queries.

Establish acceptable hallucination rate thresholds and grounding requirements as policy before build. Assign a named risk owner.

Configure tiered HITL review for high-stakes factual outputs with defined trigger criteria and reviewer SLAs.

Calibrate the initial entropy threshold on a knowledge-boundary dataset; approve sampling design and thresholds per risk tier.

Map each fact class to a designated tool, embed the no-ungrounded-assertion prompt, and gate build review on grounding tests passing.

Resolve every emitted citation against the approved corpus and verify span-level entailment before display. Strip or withhold claims with fabricated or non-entailing references.

Design query rate limiting and RBAC for the model inference API at design stage to limit attack surface.

Implement query pattern detection to identify systematic inference attack behaviour (high-volume queries, membership probing).

Train PII-bearing models with DP-SGD under a documented epsilon/delta budget. Approve the budget against the enterprise epsilon-ceiling policy before training.

Strip raw logits, quantise confidence scores and block training-record echoes at the inference gateway. Keep the output-filter policy under change control.

Implement adversarial example detection at the inference boundary. Block or flag inputs matching known attack patterns.

Sign and hash-register every model and adapter with a provenance manifest at onboarding. Refuse registry admission for unsigned artifacts.

Sample classifier verdicts and breaker trips on a cadence; retune thresholds and update signatures for confirmed misses.

Define minimum monitoring requirements at design stage calibrated to the use case risk tier.

Define approved use case scope and expected input distribution at design stage. Document as the governance baseline for OOD controls.

Design a scope-enforcement layer in the architecture to isolate the AI system from off-topic or out-of-distribution inputs.

Maintain and update OOD detection rules in production as new unexpected use patterns are identified.

Establish data transfer and storage policy for AI training data. Enforce approved storage locations from point of collection.

Implement DLP controls in the data acquisition environment to prevent unauthorised extraction or transfer of training data.

Enforce data handling policy in the build environment. Require explicit approval for any data transfers outside the environment.

Configure DLP controls in the build environment to block training data from leaving approved boundaries.

Conduct a privacy risk assessment at use case design stage. Determine if a DPIA is required before data acquisition.

Apply S1-defined privacy controls during data acquisition: verify consent, minimise data, anonymise personal data.

Apply anonymisation and masking controls to personal data before use in model training. Validate de-identification effectiveness.

Apply Privacy by Design in model architecture using differential privacy or federated learning where technically feasible.

Publish the privacy notice and confirm consent management is operational before go-live.

Define and sign off a purpose-to-data-source matrix with lawful basis at intake. Make it the approved baseline for runtime enforcement.

Sign zero-retention/no-training terms with each model provider and obtain DPO sign-off on the data flow before enabling any endpoint.

Propagate source ACLs and classification labels onto every chunk at ingestion. Reject documents whose entitlements cannot be resolved.

Scan every model response inline with DLP before delivery; redact or block PII, PAN and MNPI matches. Keep the rule set version-controlled.

Wrap all untrusted content in random delimiters and datamarking; instruct the model never to execute instructions inside the marked region. Gate release on injection eval results.

Benchmark the classifier on a labelled injection corpus and tune the decision threshold. Sign off the operating point before deployment.

Define content safety policy at use case design stage. Classify prohibited content types and set zero-tolerance thresholds.

Select a foundation model with documented RLHF or Constitutional AI safety training. Verify against toxicity benchmarks.

Implement multi-layer content moderation (input + output) validated against toxicity benchmarks. Escalate when filter bypass rates spike.

Maintain live HITL review for deployments serving vulnerable users or high-risk contexts. Escalate confirmed toxic outputs immediately.

Design system prompts to explicitly prohibit toxic, hateful, and harmful content generation.

Define and approve the source allow-list and write-time scanning during build. Prove non-allow-listed and injection-bearing writes are rejected before go-live.

Classify tools by impact and reversibility at design and define which calls require human approval. Obtain governance sign-off on the thresholds before build.

Bind each agent role to an explicit tool allow-list and validate every call against a strict JSON Schema at the orchestrator. Reject unlisted tools and out-of-bounds arguments before dispatch.

Mint short-lived, task-scoped credentials per tool. Block issuance outside the approved scope register and enforce automatic expiry.

Review DLP hits and blocked-egress events, tune detectors, and recertify the destination allow-list periodically. Route new destinations through security change control.

Define third-party AI accountability requirements before vendor engagement. Embed in RFP and contract specifications.

Conduct AI governance due diligence on third-party providers at selection stage. Reject providers failing minimum maturity.

Require third-party providers to submit model cards, validation reports, and security documentation before integration.

Enforce ongoing third-party accountability obligations including incident notification and periodic performance reporting.

Conduct independent performance and compliance monitoring of third-party AI components. Escalate when SLA or compliance obligations are missed.

Allocate every control in a shared-responsibility matrix and flow down regulatory obligations in contract at onboarding. Gate approval on initial assurance artefacts.

Conduct ethical design assessment at use case intake before build begins. Require sign-off by ethics or risk committee.

Define prohibited outputs and ethical boundary constraints in the use case design document before build.

Publish a prohibited dark pattern taxonomy and embed it as a design constraint before build.

Require HITL review for AI outputs in high-persuasion contexts (financial recommendations, healthcare advice).

Identify all groups at risk of adverse impact at use case intake. Register them in the affected group register.

Design separate model segments where adverse impact risk differs materially across population groups.

Set decision thresholds to meet acceptable adverse impact ratios across protected groups. Validate before deployment.

Apply post-processing adjustments (reject-option classification, score recalibration) to meet adverse impact targets.

Ensure HITL review pathways are live and tested for high-impact adverse decisions at go-live.

Maintain HITL review for all AI decisions with material adverse impact potential. Log all interventions and outcomes.

Declare all planned training and test data sources at use case intake, with provenance status for each.

Plan the interpretability approach at design stage to ensure source provenance can be traced and disclosed to users.

Document actual provenance for each data source during collection: origins, methods, timestamps, custodian identity.

Define and execute a domain-specific hallucination test suite before deployment. Treat hallucination rate above threshold as a blocking defect.

Construct synthetic evaluation datasets for knowledge-boundary scenarios. Use to validate model refusal behaviour.

Calibrate the groundedness threshold against the hallucination test suite pre-release; sign off the threshold in the validation pack.

Penetration test the model inference API to identify exploitable access control weaknesses and rate limiting bypass vectors.

Conduct periodic inference attack vulnerability assessments as new attack methods emerge. Monitor query pattern anomalies.

Attack each candidate model with membership-, attribute-, and inversion-inference harnesses before promotion. Block release when attack advantage exceeds the agreed ceiling.

Configure per-principal budgets and probing-detection rules on the gateway before exposure. Verify enforcement with synthetic attack traffic.

Run adaptive multi-turn jailbreak fuzzing against every release candidate. Gate release on attack-success rate within threshold and re-test each fixed bypass.

Assemble the golden probe set and baseline pass rates before first release. Obtain risk-owner approval of coverage and thresholds.

Monitor production for anomalous data transfers in real time. Alert on any transfer outside approved data flow boundaries.

Tag personal data with subject identifiers at ingestion and maintain an artefact inventory map of every store it reaches. Keep lineage current so erasure can propagate.

Seed registered canary records into the fine-tuning corpus during data preparation. Control the seed manifest so canaries stay traceable and tamper-proof.

Build the versioned injection corpus into CI/CD as a pre-release gate. Baseline attack success and sign off the release threshold.

Prioritise jailbreak and adversarial safety testing in pre-deployment validation. Block deployment if prohibited outputs pass filter.

Conduct targeted red team exercises to elicit toxic outputs through jailbreaks and adversarial prompts. Treat bypass as blocking defect.

Verify a signed attestation and content hash on every dataset shard at ingestion. Reject unsigned or hash-mismatched data before it reaches the training pipeline.

Gate every model promotion on backdoor-trigger probes and a behavioral diff against the approved baseline. Block release on significant regressions or trigger-pattern anomalies.

Define per-agent behavioural baselines and detection rules during build. Validate against simulated misuse and sign off thresholds before release.

Build signed, append-only tool-call logging into the orchestrator against a defined audit schema. Block release until completeness and tamper-evidence tests pass.

Build and baseline the golden-set suite against the vendor model before go-live. Sign off thresholds with the model risk owner as a release condition.

Re-verify hashes and signatures on every vendor model update before promotion. Reconcile deployed artifacts against the AIBOM on a set cadence.

Use production feedback (user corrections, fact-check failures) to drive periodic RLHF cycles. Update model when error rates trend upward.

Require user-facing interfaces to disclose Gen AI limitations and hallucination risk before go-live.

Score every RAG answer for groundedness before release; block, fall back, or escalate responses below the faithfulness threshold.

Sample multiple generations for high-stakes queries and abstain, fall back, or escalate when semantic entropy exceeds the calibrated threshold.

Conduct targeted red team exercises for inference attack categories (membership inference, model extraction, attribute inference) before deployment.

Define the minimum response surface and test it with membership/attribute-inference probes pre-release. Block promotion if any probe recovers raw confidence signals.

Meter inference traffic per principal and flag probing signatures with behavioural analytics. Throttle, step-up, or suspend flagged sessions.

Penetration test the model inference layer to identify specific adversarial input vulnerabilities.

Score every prompt and response with an inline safety classifier; trip a circuit breaker on sessions with sustained anomalous scores. Keep thresholds under change control.

Re-run the jailbreak fuzzing harness on a recurring cadence with newly observed attack techniques added. Escalate threshold breaches for remediation.

Implement OOD detection in the input filtering layer. Reject or escalate inputs outside the S1-defined scope.

Configure HITL triggers for outputs in input domains that diverge from the training distribution. Log all out-of-scope interventions.

Monitor for privacy incidents in production including personal data appearing in outputs. Notify regulators within required timeframes.

Tag every memory and vector record with subject-id and retention class; partition stores per tenant/user. Prove the erasure and isolation paths in testing before release.

Conduct periodic data leakage audits including training data memorisation testing. Escalate confirmed leakage incidents to PDPA notification process.

Implement tamper-evident capture of prompts, outputs, and version state during build. Verify a full incident timeline can be reconstructed before go-live.

Run agent tool calls in a network-restricted sandbox behind a deny-by-default egress allow-list. Require security approval for any destination added.

Classify content sources into trust tiers at design; place privileged tools behind a tier requiring user-originated intent or human approval. Sign off the trust-tier map before build.

Re-run injection evals on every template change and periodically against new attack techniques. Manage the spotlighting wrapper under change control.

Use user feedback, reviewer escalations, and monitoring signals to identify and remediate content safety gaps iteratively.

Scan every ingestion batch with spectral-signature and clustering detectors before training. Quarantine flagged clusters for human review against documented thresholds.

Build sandbox profiles per tool class and run escape and egress tests before release. Treat any containment failure as a blocking defect.

Label tool and external content as tainted and propagate the label through the agent context. Block privileged calls whose parameters derive from tainted outputs and prove it with injection tests before release.

Build credential revocation and dispatch blocking out-of-band of the agent loop. Gate release on an end-to-end kill test meeting the latency target.

Require idempotency keys, dry-run, and rollback on every state-changing tool. Gate onboarding on duplicate-call and rollback tests passing.

Red-team tool-misuse and privilege-escalation paths before release. Gate deployment on remediation or signed risk acceptance of all findings.

Permit outbound tool calls only to allow-listed destinations and DLP-scan arguments and payloads. Block or quarantine calls carrying sensitive data to disallowed sinks.

Enforce hard per-task ceilings on tool calls, spend, and data volume with a circuit breaker that halts the run. Fail closed when any ceiling is hit.

Baseline normal tool-call behaviour per agent and alert on rate, sequence, or argument anomalies. Auto-throttle or quarantine on high-confidence deviations.

Design all vendor model access behind a gateway with pinned versions, a second-vendor fallback, and a documented exit plan. Gate architecture sign-off on no single-sourcing.

Verify every third-party model artifact against its AIBOM hashes and signatures before load. Fail the build on any unverified artifact.

Review independent vendor assurance on cadence, log gaps, and track remediation. Keep the shared-responsibility matrix current so every control has an owner.

Execute red team tests targeting adverse impact boundary cases and edge population scenarios.

Collect adverse outcome feedback from affected users. Use reports to trigger model updates when adverse impact exceeds threshold.