AI Regulation 2026: EU AI Act, US & Global Framework

Q: What does the EU AI Act require from businesses?

The EU AI Act requires risk assessment and classification of AI systems. High-risk AI (used in healthcare, employment, law enforcement) needs conformity assessments, human oversight, transparency documentation, and data governance. General-purpose AI providers must publish training data summaries and comply with copyright law. Violations face fines up to 7% of global revenue.

Q: How does US AI regulation differ from the EU approach?

The US approach is more sector-specific and less prescriptive. Executive Orders require safety testing for frontier models (above compute thresholds), reporting to the government, and red-teaming. Sector regulators (FDA, SEC, FTC) apply existing frameworks to AI. There is no comprehensive federal AI law equivalent to the EU AI Act.

Q: Do AI regulations apply to businesses in Indonesia?

Indonesia is developing its own AI governance framework aligned with ASEAN guidelines. International businesses using AI in Indonesia should follow global best practices. If serving EU customers, EU AI Act compliance is mandatory regardless of company location. Most multinational businesses are adopting EU standards as a global baseline.

Bali Zero Editorial

AI Regulation: The Global Landscape

2026 marks the year AI regulation became serious. The EU AI Act is fully enforceable. US Executive Orders shape industry practices. And Asia is finding its own path.

For businesses operating internationally, understanding this landscape isn't optional.

The EU AI Act

Key Requirements (In Effect)

Risk-Based Classification:

Unacceptable: Banned (social scoring, subliminal manipulation)
High-Risk: Heavy regulation (hiring, credit, healthcare)
Limited Risk: Transparency required (chatbots, deepfakes)
Minimal Risk: Largely unregulated (spam filters, games)

Penalties:

Up to €35M or 7% global revenue
Graduated enforcement through 2027

What This Means for Business

AI Use Case	Risk Level	Requirements
Hiring screening	High	Impact assessment, human oversight, logging
Customer chatbot	Limited	Disclose AI nature
Document summary	Minimal	None specific
Credit scoring	High	Explainability, appeal rights
Marketing personalization	Limited	Transparency about profiling

United States Approach

The US takes a sectoral rather than comprehensive approach:

US AI Governance

Executive Order 14110 (Active):

Reporting requirements for large models
Safety testing mandates
Federal procurement standards

Sector-Specific:

FDA: AI medical devices
FTC: AI consumer protection
SEC: AI in financial services
DOL: AI in employment

State Level:

California AI Safety Bill (debated)
Various state-specific rules

China's Regulatory Framework

China has moved faster than anyone on AI regulation:

Regulation	Focus	Status
Algorithm Recommendations	Social media, e-commerce	Enforced
Deep Synthesis (Deepfakes)	Identity verification	Enforced
Generative AI Rules	Model registration, content	Enforced
Draft AI Law	Comprehensive framework	Pending

Southeast Asia Including Indonesia

Regional Status

Indonesia:

No comprehensive AI law yet
Draft Personal Data Protection (PDP) applies
Kominfo guidelines for AI use
Expected framework by 2027

Singapore:

AI Governance Framework (voluntary)
Model AI Governance Framework for Generative AI
Leading regional approach

ASEAN:

AI Governance Guidelines adopted
Interoperability efforts ongoing

Compliance Roadmap for Business

Immediate Actions (Now)

Audit AI usage — Document all AI tools and applications
Classify risk levels — Map against EU framework
Assess vendors — Ensure AI providers meet standards
Update disclosures — Inform users about AI interaction

Medium-Term (2026)

Implement logging — Maintain AI decision records
Build oversight — Human review processes
Train staff — AI literacy programs
Update contracts — AI liability clauses

Long-Term (2027+)

Certification preparation — For high-risk applications
Continuous monitoring — Regulatory change tracking
Industry collaboration — Best practice development

The Indonesia Angle

For businesses operating in Indonesia with international exposure:

EU customers? EU AI Act applies to you
US operations? Sectoral rules apply
Local only? Current flexibility, but prepare

The smart play: build to EU standards now. It's the highest bar and provides portability to any market.

Key Provisions for Common AI Uses

Application	Key Requirement	Timeline
AI Chatbots	Disclosure	Now
AI-Generated Content	Labeling	Now
AI Hiring Tools	Impact assessment	Aug 2025
AI Credit Decisions	Explainability	Aug 2025
General Purpose AI	Model cards	Aug 2025

Emerging Trends

Watch These Developments

2026 Focus Areas:

Agentic AI — Regulation for autonomous systems
AI Liability — Who's responsible when AI fails?
Open Source — Different rules for open models?
Compute Governance — Regulating training infrastructure

2027+ Horizon:

Global interoperability frameworks
AI auditing standards
Mandatory AI impact assessments

Questions?

Ask Zantara

AI-powered answers from our knowledge base

AI Regulation: The Global Landscape

2026 marks the year AI regulation became serious. The EU AI Act is fully enforceable. US Executive Orders shape industry practices. And Asia is finding its own path.

For businesses operating internationally, understanding this landscape isn't optional.

The EU AI Act

Key Requirements (In Effect)

Risk-Based Classification:

Unacceptable: Banned (social scoring, subliminal manipulation)
High-Risk: Heavy regulation (hiring, credit, healthcare)
Limited Risk: Transparency required (chatbots, deepfakes)
Minimal Risk: Largely unregulated (spam filters, games)

Penalties:

Up to €35M or 7% global revenue
Graduated enforcement through 2027

What This Means for Business

AI Use Case	Risk Level	Requirements
Hiring screening	High	Impact assessment, human oversight, logging
Customer chatbot	Limited	Disclose AI nature
Document summary	Minimal	None specific
Credit scoring	High	Explainability, appeal rights
Marketing personalization	Limited	Transparency about profiling

United States Approach

The US takes a sectoral rather than comprehensive approach:

US AI Governance

Executive Order 14110 (Active):

Reporting requirements for large models
Safety testing mandates
Federal procurement standards

Sector-Specific:

FDA: AI medical devices
FTC: AI consumer protection
SEC: AI in financial services
DOL: AI in employment

State Level:

California AI Safety Bill (debated)
Various state-specific rules

China's Regulatory Framework

China has moved faster than anyone on AI regulation:

Regulation	Focus	Status
Algorithm Recommendations	Social media, e-commerce	Enforced
Deep Synthesis (Deepfakes)	Identity verification	Enforced
Generative AI Rules	Model registration, content	Enforced
Draft AI Law	Comprehensive framework	Pending

Southeast Asia Including Indonesia

Regional Status

Indonesia:

No comprehensive AI law yet
Draft Personal Data Protection (PDP) applies
Kominfo guidelines for AI use
Expected framework by 2027

Singapore:

AI Governance Framework (voluntary)
Model AI Governance Framework for Generative AI
Leading regional approach

ASEAN:

AI Governance Guidelines adopted
Interoperability efforts ongoing

Compliance Roadmap for Business

Immediate Actions (Now)

Audit AI usage — Document all AI tools and applications
Classify risk levels — Map against EU framework
Assess vendors — Ensure AI providers meet standards
Update disclosures — Inform users about AI interaction

Medium-Term (2026)

Implement logging — Maintain AI decision records
Build oversight — Human review processes
Train staff — AI literacy programs
Update contracts — AI liability clauses

Long-Term (2027+)

Certification preparation — For high-risk applications
Continuous monitoring — Regulatory change tracking
Industry collaboration — Best practice development

The Indonesia Angle

For businesses operating in Indonesia with international exposure:

EU customers? EU AI Act applies to you
US operations? Sectoral rules apply
Local only? Current flexibility, but prepare

The smart play: build to EU standards now. It's the highest bar and provides portability to any market.

Key Provisions for Common AI Uses

Application	Key Requirement	Timeline
AI Chatbots	Disclosure	Now
AI-Generated Content	Labeling	Now
AI Hiring Tools	Impact assessment	Aug 2025
AI Credit Decisions	Explainability	Aug 2025
General Purpose AI	Model cards	Aug 2025

Emerging Trends

Watch These Developments

2026 Focus Areas:

Agentic AI — Regulation for autonomous systems
AI Liability — Who's responsible when AI fails?
Open Source — Different rules for open models?
Compute Governance — Regulating training infrastructure

2027+ Horizon:

Global interoperability frameworks
AI auditing standards
Mandatory AI impact assessments

Questions?

Ask Zantara

AI-powered answers from our knowledge base

AI Regulation 2026: The Global Framework Taking Shape

AI Regulation: The Global Landscape

The EU AI Act

What This Means for Business

United States Approach

China's Regulatory Framework

Southeast Asia Including Indonesia

Compliance Roadmap for Business

Immediate Actions (Now)

Medium-Term (2026)

Long-Term (2027+)

The Indonesia Angle

Key Provisions for Common AI Uses

Emerging Trends

Questions?

Ask Zantara

Questions about how this applies to your case?

AI Regulation 2026: The Global Framework Taking Shape

AI Regulation: The Global Landscape

The EU AI Act

What This Means for Business

United States Approach

China's Regulatory Framework

Southeast Asia Including Indonesia

Compliance Roadmap for Business

Immediate Actions (Now)

Medium-Term (2026)

Long-Term (2027+)

The Indonesia Angle

Key Provisions for Common AI Uses

Emerging Trends

Questions?

Ask Zantara

Questions about how this applies to your case?