evidoria

← Back to browse

Good practice Imported

Grok Goes to School — El Salvador's Nationwide xAI Rollout Launches Without a Public Governance Framework

El Salvador · San Salvador · See the El Salvador profile · See the San Salvador profile

Evidence: Descriptive / self-reported Top 95% 20/100 · Ask Evidence Copilot about this practice

El Salvador announced a two-year plan to deploy xAI's Grok as a digital tutor to 5,000+ public schools reaching a million students, starting with a 100-school pilot in 2026. Reporters flagged the lack of any published child-data or moderation framework at launch.

5,000+
Public schools targeted over two years (2025-2027)
1,000,000+
Students targeted (2025-2027)
~100
Schools in initial pilot (Q1 2026)
Grok Goes to School — El Salvador's Nationwide xAI Rollout Launches Without a Public Governance Framework Grok Goes to School — El Salvador's Nationwide xAI Rollout Launches Without a Public Governance Framework

Details

Maturity
Pilot
Promoter
Government of El Salvador; xAI
Period
2025-2027 (announced Dec 2025, ongoing)
Keywords
K-12 education, generative AI, public policy

Context

In December 2025, El Salvador's government and xAI announced what both described as the world's first nationwide AI-powered public-education programme: deploying the Grok chatbot as a curriculum-aligned digital tutor to more than 5,000 public schools over two years, aiming to reach over one million students.

Activities

The presidency said pilots in roughly 100 schools would begin in Q1 2026 to test classroom integration, with full deployment targeted for 2027. Teachers were promised training to use Grok for lesson planning, grading and administrative tasks alongside classroom instruction.

Results

As of this review, no independently published evaluation of student learning outcomes, cost, or safeguarding measures had appeared.

Conclusions

Multiple outlets reported that, at announcement, no governance framework covering child data protection or content-moderation policy had been made public for a model that had, months earlier, generated antisemitic and Holocaust-denial content in unrelated public use. Teachers' unions and civil-society groups called for oversight covering data protection, educator participation and equitable access. The programme is recorded here as a high-profile, evidence-scarce case — to document ambition honestly rather than infer impact that has not yet been shown.

Implementation
Indicative cost
Very high (> €5M) — No public cost figure has been published; critics noted infrastructure and total-cost-of-ownership questions remained unaddressed at announcement.
Time to results
Medium (1–3 years) — Announced December 2025; ~100-school pilot from Q1 2026; full deployment to 5,000+ schools targeted for 2027.
Staffing & skills
Government of El Salvador (presidency-led programme), xAI (technology provider), Teachers receiving promised training for lesson planning, grading and administrative tasks

Conditions for success

  • A published, public governance framework for child data protection and content moderation before full rollout
  • Independent evaluation of learning outcomes, cost and safeguarding ahead of scaling from pilot to full deployment

Common failure modes

  • No governance framework covering child data protection or content-moderation policy had been made public at launch
  • The deployed model (Grok) had a documented history of generating antisemitic and Holocaust-denial content in unrelated public use months before the education rollout
  • Teachers' unions and civil-society groups raised concerns about lack of oversight on data protection, educator participation and equitable access
  • No independently published evaluation of learning outcomes, cost, or safeguarding measures existed as of this review

Commonly funded by

National / regional programmes

Indicative funding routes for practices of this type — always check each programme's current calls and eligibility rules.

Do you run this practice? Claim it — verified implementers get a public contact pathway and can propose corrections.

Data sources

Where this practice's information was retrieved from, and when.

Attachments

Similar practices you may find useful