OpenAI · OpenAI GPT-5 System Card · View original document ↗

Refusal Training and Output Mitigation Measures

Medium severity Medium confidence Explicitdocumentlanguage Unique · 0 of 352 platforms
Share 𝕏 Share in Share 🔒 PDF
Recent governance activity OpenAI recorded 14 documented changes in the last 30 days.
Start monitoring updates
Monitor governance changes for OpenAI Create a free account to receive the weekly governance digest and monitor one platform for governance changes.
Create free account No credit card required.
Document Record

What it is

The card discloses that GPT-5's training includes behavioral mitigations in the form of refusal responses and output constraints corresponding to the risk categories identified in the Preparedness Framework evaluation, applied as a condition of deployment authorization.

This analysis describes what OpenAI's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology

ConductAtlas Analysis

Why it matters (compliance & governance perspective)

This provision establishes that certain categories of output are constrained at the training level rather than solely through operator or user controls, and the scope and reliability of these constraints are directly relevant to deployers' own risk assessments and regulatory compliance representations.

Interpretive note: The robustness benchmarks, bypass rates, and update mechanisms for refusal training are not specified in the available document text, limiting assessment of the reliability of this mitigation.

Consumer impact (what this means for users)

Under these terms, GPT-5 is trained to refuse certain categories of requests related to CBRN, cyberoffense, and other identified risk categories, and these refusal behaviors apply across all access tiers and operator configurations within the bounds of OpenAI's policy.

Cross-platform context

See how other platforms handle Refusal Training and Output Mitigation Measures and similar clauses.

Compare across platforms →

Monitoring

OpenAI has changed this document before.

Receive same-day alerts, structured change summaries, and monitoring for up to 25 platforms.

Get Monitor Or create a free account →
▸ View Original Clause Language DOCUMENT RECORD
"
GPT-5 incorporates refusal training and output constraints applied prior to deployment to mitigate risks identified in Preparedness Framework evaluations across CBRN, cybersecurity, persuasion, and model autonomy categories.

— Excerpt from OpenAI's OpenAI GPT-5 System Card

ConductAtlas Analysis

Institutional analysis (Compliance & governance intelligence)

1) REGULATORY LANDSCAPE: Trained refusal and output constraint measures engage the EU AI Act's technical robustness and safety requirements for high-risk and GPAI systems, as well as the FTC's standards for AI systems that make safety-related representations. The reliability and consistency of refusal behaviors may be subject to scrutiny under consumer protection frameworks if they are found to be inconsistently applied. 2) GOVERNANCE EXPOSURE: Medium. The card discloses that refusal training was applied but does not specify the robustness benchmarks used to validate refusal reliability, the rate of refusal bypass observed in red-teaming, or the mechanism by which refusal behaviors are updated when new elicitation methods are identified. This creates uncertainty for deployers who rely on these mitigations in their own risk documentation. 3) JURISDICTION FLAGS: EU deployers must assess whether trained refusal behaviors constitute sufficient technical and organizational measures under the AI Act and GDPR. Deployers in healthcare or financial services must assess whether refusal behaviors adequately address sector-specific harm categories beyond those disclosed in the card. 4) CONTRACT AND VENDOR IMPLICATIONS: Enterprise agreements should specify OpenAI's obligations to maintain and update refusal training in response to newly identified elicitation techniques, and whether OpenAI provides notification to operators when significant changes to refusal behavior are made. 5) COMPLIANCE CONSIDERATIONS: Compliance teams should not treat disclosed refusal training as a substitute for their own application-level content filtering, particularly in deployments serving vulnerable populations or operating in regulated sectors where the consequences of refusal failure are material.

Full compliance analysis

Regulatory citations, enforcement risk, and due diligence action items.

Track 3 platforms — free Get Monitor

Free: track 3 platforms + weekly digest. Monitor: 25 platforms + same-day alerts. No credit card required.

Applicable agencies

  • FTC
    FTC authority over unfair or deceptive practices in AI systems may be relevant if refusal training representations are found to materially overstate the reliability of output constraints.
    File a complaint →

Provision details

Document information
Document
OpenAI GPT-5 System Card
Entity
OpenAI
Document last updated
July 6, 2026
Tracking information
First tracked
July 6, 2026
Last verified
July 6, 2026
Record ID
CA-P-013392
Document ID
CA-D-00923
Evidence Provenance
Source URL
Wayback Machine
Content hash (SHA-256)
3ccccb27d7c63ef56fc7aca27175534c6244c9eab5e10f1f78b0ed9ac9909e4f
Analysis generated
July 6, 2026 22:05 UTC
Methodology
Evidence
✓ Snapshot stored   ✓ Hash verified
Citation Record
Entity: OpenAI
Document: OpenAI GPT-5 System Card
Record ID: CA-P-013392
Captured: 2026-07-06 22:05:59 UTC
SHA-256: 3ccccb27d7c63ef5…
URL: https://conductatlas.com/platform/openai/openai-gpt-5-system-card/refusal-training-and-output-mitigation-measures/
Accessed: July 7, 2026
Permanent archival reference. Stable identifier suitable for legal filings, compliance documentation, and research citation.
Classification
Severity
Medium
Categories

Other risks in this policy

Compliance Governance Intelligence

Need to monitor specific governance provisions?

Compliance includes provision-level monitoring, governance timelines, regulatory mapping, and audit-ready analysis.

Arbitration clauses AI governance Data rights Indemnification Retention policies
Get Compliance

Or start with Monitor →

Built from archived source documents, structured governance mappings, and historical version tracking.

Frequently Asked Questions

What does OpenAI's Refusal Training and Output Mitigation Measures clause do?

This provision establishes that certain categories of output are constrained at the training level rather than solely through operator or user controls, and the scope and reliability of these constraints are directly relevant to deployers' own risk assessments and regulatory compliance representations.

How does this clause affect you?

Under these terms, GPT-5 is trained to refuse certain categories of requests related to CBRN, cyberoffense, and other identified risk categories, and these refusal behaviors apply across all access tiers and operator configurations within the bounds of OpenAI's policy.

Is ConductAtlas affiliated with OpenAI?

No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by OpenAI.