AI Accelerates Workforce Restructuring and Regulatory Pressure Ahead of AI Act

On March 31, 2026, Larry Ellison initiated the largest restructuring^[1] in Oracle’s history. According to TD Cowen estimates, the company is cutting between 20,000 and 30,000 jobs^[2], about 18% of its 162,000 global employees. The layoffs affect areas including Revenue & Health Sciences and SaaS & Virtual Operations Services, where reductions exceed 30%, as well as^[5] the Net. Suite India team. The restructuring program’s cost in fiscal 2026 may reach $2.1 billion^[3], yet it aims to free up $8–10 billion in annual cash flow^[4] to build AI data center infrastructure, with committed investments around $35 billion. Meanwhile, Oracle’s stock price, despite a 27% decline since the year’s start, jumped 5% after the announcement.

Record Layoffs Driven by AI

The wave of layoffs doesn’t stop at Oracle^[6]. Independent tracker RationalFX data shows that between January 1 and the end of March 2026, the tech industry eliminated 45,363 jobs^[7], of which were directly linked to AI deployments and automation^[8]. Amazon cut approximately 16,000 positions^[9], Atlassian 1,600 (around 10% of its workforce), Wise. Tech Global 2,000, and Block, led by Jack Dorsey, 4,000 employees. Epic Games reduced about 1,000 jobs, e. Bay 800, and Meta around 700 in the second round of cuts. RationalFX forecasts that if this trend continues, up to 264,730 tech jobs^[10] will disappear in 2026, surpassing 245,000 lost in 2025.

Freedom of Speech Dispute and AI Regulation

On March 27, 2026, federal judge Rita F. Lin from San Francisco issued a temporary ban^[11] on enforcing President Donald Trump’s directive dated February 27, 2026, which ordered all federal agencies to stop using Anthropic technology^[12]. The company, led by Dario Amodei, argued the decision violated the Administrative Procedure Act^[13] and the First Amendment because it was retaliation for its public stance on AI ethics and safety. The judge ruled the administration’s actions as ‘classic unlawful retaliation’^[14] and noted that evidence suggests an ideological motive rather than a genuine national security threat. The court delayed the order by 7 days^[15] to allow the government to file an appeal, but the ruling sets a precedent protecting AI companies from political pressure.

Agent Safety as the New Standard

At the same time, the federal government in Washington is trying to organize its regulatory approach^[16]. On March 20, 2026, Donald Trump presented the ‘National AI Legislative Framework’^[17], a White House document designed to guide Congress toward adopting a single nationwide AI law, overriding state rules such as the California AI Safety Bill and the Texas AI Act. Michael Kratsios emphasized that ‘we need a single national AI framework’^[18]. The framework includes six policy goals^[19], including child protection, privacy rights, freedom of speech, and building U. S. AI leadership, but does not plan to create a specialized oversight agency, unlike the European strategy based on the AI Act.

Boom in AI Agents in Business

On the technology front, on April 3, 2026, Microsoft released the open-source Agent Governance Toolkit^[20]—a complete set of tools for supervising autonomous AI agents. The core of the platform, Agent OS, enforces policies with a delay under 0.1 ms^[21] and supports rule definition languages such as YAML, OPA Rego, and Cedar. Agent Mesh assigns agents cryptographic identities (Ed25519) and trust scores^[22] on a 0–1000 scale across five levels. Agent Runtime implements permission rings borrowed from processor architecture^[23] and an emergency kill switch, while the Agent Compliance module maps deployments to EU AI Act, HIPAA, and SOC2 requirements^[24], collecting audit evidence. A staggering 97% of companies expect a serious agent-related incident^[25] in 2026, while Cisco’s analysis revealed that 26% of agent platform plugins contain^[26] at least one vulnerability.

Poland on the AI Act Fast Track

The business potential of AI agents is confirmed by Salesforce’s results. In its Q4 fiscal 2026 report, the company stated that its Agentforce platform reached $800 million in annual recurring revenue^[27], growing 169% year-over-year in just 15 months on the market^[28] and gaining 29,000 paying customers. Combined with the Data 360 suite, AI product revenues hit $2.9 billion ARR^[29], increasing about 200% year-over-year, and every one of the quarter’s top 10 contracts featured an Agentforce component. At the same time, total group revenue grew to $41.5 billion annually^[30], and the share repurchase program reached $50 billion^[31]. However, the CFO warned that the rapid ‘commoditization’ of token pricing could squeeze margins^[32] and force a shift from a simple per-conversation model to unique IP-based offerings.

Amid these global changes, Poland enters a crucial phase of adapting to European law. The Ministry of Digitization, led by Krzysztof Gawkowski, announced priorities for 2026–2027^[33]: building two ‘AI factories’ in Poznań and Kraków—centers with supercomputers, data centers, and training programs—implementing the EU AI Act, and launching regulatory sandboxes. According to the European Commission’s timeline, each member state must run at least one sandbox^[34] for AI systems starting August 2, 2026, and most provisions of the regulation will begin enforcement on that date^[35]. At the national level, the ministry plans an AI systems law and the establishment of a Commission for AI Development and Safety but has not yet presented a bill or schedule. Meanwhile, the AI Act rules, with penalties reaching €35 million or 7% of global turnover, will apply directly, forcing Polish companies and administration to promptly classify^[36] system risks and build compliance documentation.

Related posts:

• Breakthrough Week in AI: China’s Offensive, Agent Platforms, and Security Crises
• Iran–Israel–USA Conflict: Status as of March 2, 2026
• NVIDIA, New Regulations, and a Wave of Funding: The Week AI Agents Enter the Mainstream
• AI Agents Enter Offices, Attacks Surge, and AI Act Pressure Intensifies