URGENT June 25, 2026
✍️ Carlos Montiel
⏱ 11 min read
The EU AI Act reached full enforcement on August 2, 2026. If your company sells to European customers, processes EU citizens' data, or uses AI systems that affect people in Europe — this law applies to you even if you're based in Guatemala. Fines of up to €35M or 7% of your global revenue.
⚠️ Deadline: August 2, 2026
This is the full-enforcement date of the EU AI Act (EU Regulation 2024/1689). High-risk AI systems that aren't compliant by that date are exposed to regulatory sanctions in every EU member state.
What Is the EU AI Act?
The EU AI Act (EU Regulation 2024/1689) is the world's first comprehensive legal framework on artificial intelligence. It entered into force on August 1, 2024 and has a gradual implementation period culminating in full enforcement on August 2, 2026.
The core principle: it classifies AI systems by risk level and applies proportional requirements. Not all AI is banned or requires the same level of compliance — but ignoring it entirely isn't an option for companies with European exposure either.
Does It Apply to Companies in Guatemala and LatAm?
Yes, if any of these conditions apply to your company:
- You offer products or services to users within the European Union
- Your AI systems' outputs affect people within the EU
- You process personal data of European citizens
- You have customers or contracts with European companies that must themselves comply
Many Guatemalan export companies, BPO services, software houses with European clients, and SaaS startups fall within scope without knowing it.
Risk Category System
| Category | Examples | Consequence |
|---|
| PROHIBITED | Government social scoring, subliminal manipulation, mass biometric surveillance | Total ban — cannot be operated |
| HIGH RISK | AI in recruitment, credit/insurance, medical diagnosis, migration, judicial decisions | Strict requirements before August 2, 2026 |
| Limited risk | Chatbots, image generation, deepfakes | Transparency obligation (must disclose it's AI) |
| Minimal risk | Spam filters, AI-powered games, content recommendations | No special requirements — just good practice |
Requirements for High-Risk Systems
If your company operates high-risk AI systems (recruitment, credit, healthcare), before August 2, 2026 you must have:
- Conformity assessment: A formal audit certifying the system meets AI Act requirements
- Complete technical documentation: System manual, training data, architecture, performance metrics
- Registration in the EU database: High-risk systems must be registered in the official repository
- Risk management system: An ongoing process — not a one-time checkbox
- Mandatory human oversight: Mechanisms for humans to supervise, correct, and stop the system
- Post-deployment monitoring: Reporting serious incidents to competent authorities
Fines and Penalties
| Type of infraction | Maximum fine |
|---|
| Prohibited AI practices (social scoring, manipulation) | €35M or 7% of annual global revenue |
| Non-compliance with high-risk system requirements | €15M or 3% of annual global revenue |
| Providing incorrect information to regulatory authorities | €7.5M or 1.5% of annual global revenue |
Fines are applied by national supervisory authorities in each member state. For LatAm companies, actual exposure depends on whether they have operations, contracts, or European citizens' data — but the legal risk exists regardless of where headquarters are located.
Practical Action Plan: 6 Steps
- Inventory: List all your AI systems — chatbots, decision tools, predictive analytics, any ML model in production
- Risk classification: For each system, determine whether it's prohibited, high-risk, or limited under the AI Act's categories
- Gap analysis: Compare each high-risk system's current state against AI Act requirements
- Compliance roadmap: A plan of concrete actions with dates, owners, and budget
- Legal review: A lawyer specialized in EU law should review your contracts with European clients and your AI usage policies
- Ongoing governance: The AI Act isn't a one-time project — it's a permanent AI governance process
Is your company ready for the EU AI Act?
Carlos Montiel conducts AI compliance audits for companies in Guatemala and Latin America with European exposure. Risk assessment, technical documentation, and compliance roadmap.
Request an AI Act audit
Carlos Montiel
Enterprise AI Solutions Architect · guatemalia.com
Specialist in enterprise AI and compliance for companies in Guatemala and Latin America. Contact: guatemalia.com/en/#contact