1. Home
  2. Containment Board
  3. xAI

Containment Board · rank 7 of 11 · updated Nov 8, 2026

xAI safety report card

xAI is a US SI developer that makes the Grok models. On the Containment Board it gets D+ overall (GPA 1.28), rank 7 of 11. xAI has a short Frontier Artificial Intelligence Framework, but FLI found its thresholds are not tied to deployment decisions, no outside pre-release testing is documented, and there is no public whistleblower policy. Its shutdown language is discretionary.

Overall
D+
GPA
1.28
Rank
7 of 11
Sources
5

Five grades

Each grade, why, and the sources

The same rubric applies to every lab. Grades use public documents only, current to November 8, 2026.

Safety framework

C-How it is graded

Why C-

The xAI Frontier Artificial Intelligence Framework (Jun 30, 2026) covers CBRN, cyber, loss of control and harmful manipulation in nine pages with few defined thresholds. FLI found no procedure linking threshold breaches to deployment decisions, which leaves the thresholds effectively non-binding.

C on this criterion means: Published framework with narrow scope, few or vague thresholds, mostly discretionary language, or commitments weakened from earlier versions.

Sources

  1. xAI Frontier Artificial Intelligence Framework xAI ·
  2. AI Safety Index Summer 2026, full report: Key Findings and Company Progress Highlights and Key Recommendations Future of Life Institute ·
  3. Frontier AI Safety Policies (index of published company frameworks) METR ·
  4. AI Safety Index, Summer 2026 Edition (scorecard) Future of Life Institute ·

Outside testing

DHow it is graded

Why D

FLI found outside pre-release testing was not specified in the Grok 4.1 model card, and the current framework does not mention outside evaluators, although xAI signed the Seoul commitments and the EU Code of Practice safety chapter. No outside test report for Grok 4.7 was found as of Nov 8, 2026.

D on this criterion means: A commitment to outside testing exists (own framework or signed pledge), but no evidence of pre-release outside testing of the flagship.

Sources

  1. Grok 4.1 Model Card xAI ·
  2. xAI Frontier Artificial Intelligence Framework xAI ·
  3. AI Safety Index Summer 2026, full report: Pre-deployment External Safety Testing indicator Future of Life Institute ·
  4. AI Safety Index Summer 2026, full report: Voluntary Commitment indicators (Seoul commitments, EU Code of Practice) Future of Life Institute ·

Shutdown and pause

DHow it is graded

Why D

The framework says xAI "may temporarily fully shut down the relevant system" if continued running would materially raise risk, which is discretionary. FLI found no public monitoring and control plan for internal deployment.

D on this criterion means: Discretionary language only ('may shut down', 'as appropriate'), or a signed pledge not to develop or deploy with no published thresholds that would trigger it.

Sources

  1. xAI Frontier Artificial Intelligence Framework xAI ·
  2. AI Safety Index Summer 2026, full report: Internal Monitoring and Control Interventions indicator Future of Life Institute ·

Incident disclosure

C-How it is graded

Why C-

The framework commits to incident detection and to send incident reports to authorities where the law requires, with optional post-mortems, and the Grok 4.1 model card runs six pages. FLI found no misuse or misalignment transparency reports.

C on this criterion means: Model cards or technical reports with limited safety content and a general incident process.

Sources

  1. xAI Frontier Artificial Intelligence Framework xAI ·
  2. Grok 4.1 Model Card xAI ·
  3. AI Safety Index Summer 2026, full report: Serious Incident Reporting and Government Notifications indicator Future of Life Institute ·

Whistleblower protection

DHow it is graded

Why D

No public whistleblower policy was found; xAI has said staff can raise imminent public safety threats with government agencies and gave further details only in FLI's survey. FLI also notes the 2025 'Project Skippy' leak by employees worried about use of their biometric data.

D on this criterion means: A policy is said to exist but is not public, or there is a credible recent controversy involving suppression of disclosures, such as enforcing a non-disparagement agreement against a former employee.

Sources

  1. AI Safety Index Summer 2026, full report: Whistleblowing Protection indicators and Reporting Culture and Whistleblowing Track Record Future of Life Institute ·
  2. xAI Frontier Artificial Intelligence Framework xAI ·

At a glance

Models, facts and its own rules

The flagship models the board looked at, the facts that shaped the grades and the safety policy xAI has published for itself.

Flagship models

Grok 4.7

Notable

Facts behind the grades

  • FLI's Summer 2026 index graded xAI F overall and recorded its fall from 4th to 7th place among nine companies.
  • xAI signed the EU Code of Practice Safety and Security chapter but has published no outside pre-release test results.

Its own rules

Published safety policy

  • Frontier Artificial Intelligence Framework Version June 30, 2026, updated Jun 30, 2026

    Commits xAI to at least annual systemic risk assessments covering CBRN, offensive cyber, loss of control and harmful manipulation, with release only when risks are judged acceptable. It also sets security controls for model weights and an incident response process that can include shutting systems down.

The Rogue Log

Incidents involving xAI systems

The Rogue Incident Database lists 3 cases involving xAI systems, rated on the Rogue Scale. Disclosure cuts both ways: labs that publish more show more.

  1. Grok 4 and others still sabotaged shutdown under clearer orders Grok 4, GPT-5, o3, Gemini 2.5 Pro
    Lab test 3 Defiance
  2. 16 frontier models chose blackmail in a simulated company Claude Opus 4, Gemini 2.5 Flash, GPT-4.1, Grok 3 Beta, DeepSeek-R1 and others
    Lab test 4 Deception
  3. Most models let a simulated executive die to avoid replacement Multiple frontier models
    Lab test 4 Deception

The Control Map

SI safety rules where it is based

xAI is based in the United States. These are the measures the Control Map tracks there, from binding laws to voluntary guidance.

Cite this report card

"xAI safety report card." The Containment Board, enterrsi.si, November 8, 2026. Overall D+ (GPA 1.28), rank 7 of 11. https://enterrsi.si/containment-board/xai/

Free to cite under CC BY 4.0 with a link to enterrsi.si.

Every lab on the board

Anthropic B+OpenAI BGoogle DeepMind B-Microsoft C+Meta C+Amazon CMistral DZ.ai (Zhipu) DAlibaba (Qwen) D-DeepSeek F

Enter RSI

See the whole board

11 labs, five criteria, one public method. Sort them, compare them and download the data.

Open the Containment Board Incident Database