Global Tech Policy Tracker — 2026-04-20

Top Story

Germany's Merz Pushes for Lighter EU Industrial AI Rules — Just Before August Deadline

In a significant political intervention, German Chancellor Friedrich Merz stated on Sunday, April 19, that artificial intelligence used for industrial purposes should face less stringent EU regulation than AI in consumer-facing domains such as healthcare or law enforcement. Merz argued that more regulatory freedom was needed to boost productivity in German industry, driving a wedge into the EU's unified approach to the AI Act.

The intervention comes at a particularly sensitive moment. The EU's AI Act is entering its most demanding phase: the ban on prohibited AI practices has been in force since February 2025, General Purpose AI (GPAI) model rules applied from August 2025, but the obligations for high-risk AI systems — covering biometric identification, health, utilities, creditworthiness, and law enforcement — are scheduled to kick in August 2, 2026, just months away. The European Commission's "Digital Omnibus" proposal, agreed to in principle by the EU Council in March, would delay these high-risk rules to December 2027, but that proposal still faces parliamentary scrutiny.

Merz's statement reflects a broader tension: the EU's largest economy wants AI as a productivity tool, while Brussels insists on a risk-based framework. For companies operating in Europe, the statement sends a mixed signal — pressure from Germany's most powerful political office could delay or soften enforcement, but the legal baseline remains in place until changed. Businesses relying on AI in manufacturing, logistics, or industrial IoT should watch whether Germany translates this political stance into formal lobbying or legislative amendments in the coming weeks.

New Legislation & Regulatory Actions

United States: State AI Law Activity — Nebraska, Maryland, Maine

• What happened: State legislatures continue to pass AI-specific bills independent of federal direction. Nebraska passed a chatbot disclosure bill, Maryland enacted AI-related pricing transparency legislation, and Maine passed an AI health bill — all in the week ending April 13, 2026.
• Who it affects: Companies operating consumer-facing AI services, AI-driven pricing tools, and health-related AI systems in these states.
• Status: Enacted at the state level; details on enforcement timelines vary by state.
• Why it matters: State-level AI activity is accelerating even as the White House pushes a preemption framework designed to consolidate federal authority. This creates an increasingly complex patchwork that businesses operating in multiple states must navigate. Senator Edward Markey has introduced the "States' Right to Regulate AI Act" to block the White House executive order, calling it "lawless."

EU: Ireland's AI Enforcement Architecture Taking Shape

• What happened: Ireland's AI enforcement plans are solidifying as a national competent authority, with compliance expectations clarifying for companies domiciled there — which includes many of the world's largest tech firms given Ireland's role as a European HQ hub.
• Who it affects: Multinational tech companies with European headquarters in Ireland, including those providing GPAI models and high-risk AI applications.
• Status: Under active development ahead of the August 2026 high-risk AI system deadline.
• Why it matters: Ireland's enforcement stance will disproportionately affect Big Tech. Companies that have registered European entities there must now plan for both Irish national enforcement and EU-level oversight from the AI Office in Brussels.

1 sources

idaireland.com

idaireland.com

EU: Digital Omnibus Council Position on AI Act Simplification

• What happened: The EU Council agreed in March 2026 to a position on the "Digital Omnibus" — a sweeping simplification package that would delay high-risk AI compliance requirements from August 2026 to December 2027 and reduce obligations for certain AI system providers.
• Who it affects: Operators and deployers of high-risk AI systems across the EU, including in biometrics, credit scoring, healthcare, and law enforcement contexts.
• Status: Council position agreed; still requires European Parliament negotiations and trilogue. Timeline remains uncertain.
• Why it matters: The gap between a Council position and a final vote means the August 2026 deadline is technically still in force. Companies that bank on the delay do so at legal risk. Amnesty International and civil society groups have sharply criticized the simplification as "rolling back rights."

Enforcement & Penalties

• EU AI Office / Member States → High-Risk AI Operators: The EU AI Act's penalty regime is now technically active for prohibited practices (Article 5 violations), with fines of up to €35 million or 7% of global annual turnover — whichever is higher. For high-risk system violations, the ceiling is €15 million or 3% of turnover. The Digital Omnibus delay, if enacted, would push many enforcement windows to December 2027, but legal exposure exists today for prohibited practices.

• EU Member States → GPAI Model Providers: Since August 2025, General Purpose AI model providers face transparency, copyright, and incident reporting obligations. The EU AI Office has been building its enforcement capacity, and Ireland's emerging national framework suggests the first significant enforcement actions against GPAI providers could materialize within 2026.

Industry Response

• German Government / Chancellor Merz: In a direct political intervention, Merz called for a regulatory carve-out for industrial AI applications within the EU framework. This is notable because Germany had previously been broadly supportive of the AI Act — the shift suggests that competitive pressure from U.S. and Chinese AI deployment is reshaping even core EU member state positions.

• Senator Edward Markey (U.S. Senate): In a legislative countermove, Markey introduced the "States' Right to Regulate AI Act" in response to the White House's Executive Order 14365, which sought federal preemption of state AI laws. Markey labeled the presidential order "lawless," escalating the federal-vs-state AI governance conflict into a congressional battleground.

• Legal and Compliance Industry: Compliance advisors and legal technology firms are reporting a surge in demand for EU AI Act readiness assessments ahead of the August 2026 deadline. The growing "AI washing" enforcement risk — where companies falsely claim AI capabilities — has been flagged as an emerging area of SEC scrutiny for public companies in the U.S. as well.

Region Scorecard

Analysis: What This Means

• For EU-based AI operators: Do not assume the Digital Omnibus delay will pass before August 2026 — it almost certainly won't. Treat the August 2, 2026 high-risk AI compliance deadline as live, particularly for biometric, health, and financial AI applications. Germany's political push may eventually soften implementation, but legal liability accumulates in the interim.

• For U.S. companies deploying AI across states: The Nebraska, Maryland, and Maine bills underscore that state-level compliance is not going away, regardless of federal preemption ambitions. Build state-by-state tracking into your legal function now; consolidate compliance posture around the most demanding state requirements as a baseline.

• For AI developers and startups: The EU AI Act's GPAI obligations (in force since August 2025) mean that foundation model developers — even those outside the EU — face transparency and copyright documentation obligations if their models are accessed in Europe. Review your GPAI compliance posture against the AI Office's published guidance.

• For public companies (especially U.S.-listed): SEC scrutiny of "AI washing" is intensifying. Companies making AI capability claims in investor materials face mounting liability risk. Ensure AI-related disclosures are accurate, operationally grounded, and reviewed by legal before publication.

What to Watch Next Week

1. EU Parliament Response to Digital Omnibus: Watch for any preliminary signals from MEPs on the timeline and appetite for the simplification package — this will determine whether the August 2026 high-risk AI deadline holds or moves to 2027.

2. U.S. Senate Action on States' Right to Regulate AI Act: Senator Markey's bill is unlikely to advance quickly, but committee hearings or co-sponsorship announcements would signal the seriousness of congressional pushback against White House AI preemption strategy.

3. Ireland AI Office Announcements: Any regulatory guidance, staffing announcements, or consultation launches from Ireland's national AI competent authority would be an early indicator of enforcement priorities for GPAI model providers headquartered in Dublin.