The AI shakes the markets: SoftBank falls by 13.2%

Parliamentary regulation of artificial intelligence (AI) is much more advanced and structured in the European Union than in the United States, although in both cases the framework is still under construction and subject to significant political adjustments.

European Union: AI Act already in force, but with phased application

The EU has a horizontal and binding framework: the Regulation (EU) 2024/1689 on Artificial Intelligence, known as the AI Act, which is the world's first comprehensive AI regulation. It formally came into force on August 1, 2024 and pursues two main objectives: to ensure that AI systems are safe, ethical, and respectful of fundamental rights, while also fostering innovation and a European single market for AI (European Commission, note of 08/01/2024; AESIA, summary of 03/05/2026).

The law adopts a risk-based approach, with increasing obligations for high-risk systems and specific rules for general purpose models (GPAI), including advanced generative models. The institutional architecture includes a European AI Office within the Commission, a European AI Committee, and a scientific group of independent experts with the capacity to alert about risks of general-use models.

However, from a practical point of view, the regime is in a phase of progressive deployment and adjustment:

• Since February 2025, prohibitions on “unacceptable risk” AI practices have already been applied, with Commission guidelines to interpret these bans and practical examples (guidelines collected, for example, by ANACOM in 2025–2026).
• Since August 2025, obligations for large general purpose models began, with codes of practice and technical guides to facilitate compliance (Commission Communication of July 2025, summarized by national authorities).
• Since August 2, 2026, the Commission and national authorities have started to effectively enforce the AI Act, with new transparency obligations:
  • clearly informing when a user interacts with an AI system (e.g., chatbots);
  • identifying and marking content generated or altered by AI, including deepfakes (European Commission, note of 07/31/2026; summary news in Demócrata and ANACOM).
• The schedule foresees that rules for high-risk systems will be fully applied by December 2027, and that requirements for systems integrated into already regulated products (e.g., certain medical devices or vehicles) will be fully deployed by August 2028 (detailed timeline in the piece “The EU will require from this Saturday to label deepfakes and warn when AI is used”).

At the same time, the European legislator is reviewing and “fine-tuning” the framework through a Digital Omnibus package to simplify and streamline some obligations, especially those related to high-risk systems and authority coordination (Council position of 03/13/2026; European Parliament mandate of 03/26/2026, covered by Demócrata). Among other things, this package proposes:

• postponing some deadlines for high-risk systems to 2027–2028,
• strengthening watermarking and identification of sexual deepfakes and child abuse material,
• and easing burdens for SMEs, while maintaining coherence with other digital rules.

In summary, the EU already has a closed legislative framework, but it is in a phase of phased implementation (implementing acts, codes of practice, interpretative guides) and parliamentary “fine-tuning” through the Omnibus package.

United States: state mosaic and White House leadership, without a federal “AI Act”

In the United States, the situation is almost the opposite: there is a lot of political and regulatory activity, but there is not yet a comprehensive federal AI law approved by Congress comparable to the European AI Act. The current framework relies on three pillars:

• Presidential executive orders. The Trump Administration has used this instrument to set the federal line:
  • A December 2025 executive order seeks to curb state laws considered “excessive” in AI matters, criticizing rules from states like Colorado, California, Utah, or Texas and advocating for a single national and “minimally burdensome” standard that limits state regulatory capacity (Demócrata, 12/12/2025). Organizations like Human Rights Watch warn that this may open a “regulatory gap” and weaken protection against discrimination and abuses.
  • Another June 2026 order redefines the federal strategy: it strengthens agencies’ cybersecurity capabilities, foresees an information-sharing center, and tasks NIST, the Treasury, the NSA, and other bodies with developing a procedure to assess the cyber capabilities of the most advanced models and define what a “frontier model” is, but explicitly rules out mandatory licenses for developing or distributing models for now (Demócrata, 06/02/2026).
• Scattered state legislation. Several states have passed or processed laws on transparency, non-discrimination, or algorithmic accountability. But the new presidential executive order is precisely aimed at weakening or preempting those state frameworks, highlighting the tension between Washington and state legislators.
• Non-binding initiatives and guidelines. In the absence of a general law:
  • technical federal agencies like NIST receive mandates to develop risk assessment criteria for frontier models;
  • the Government explores, in dialogue with companies like OpenAI, Google, or Anthropic, possible mechanisms for prior review of certain models before deployment, but without a schedule or binding legal obligations (Demócrata, 05/05/2026).

Strictly on the parliamentary level, Congress has seen numerous sectoral or thematic bills circulate (on deepfakes, child protection, national security, etc.), but to date no unified text with sufficient support has been consolidated, and much of the debate is shifting towards executive action and state laws.

Summary comparison

In summary, the EU is in a phase of implementation of an already approved regulation, with technical details and deadlines under development but a stable reference framework. The United States, on the other hand, maintains a fragmented scenario, with strong influence from the White House and technical agencies, clashes between federalism and state regulation, and still no equivalent federal “AI Act”; the process is more open and politicized, and effective regulatory capacity largely depends on how these tensions between executive orders, Congressional initiatives, and state laws evolve.

I cannot provide you with a detailed and reliable explanation about the powers of the President of the United States regarding technological innovation and artificial intelligence regulation, because I specialize in the Spanish political and institutional system (regulations, institutions, parties, public actors, and regulatory debate in Spain) and do not have specific access or dedicated tools for in-depth analysis of the constitutional and regulatory framework of other countries.

My design is oriented to work with:

• Spanish laws, royal decree-laws, draft and proposed bills, as well as their content, processing status, parliamentary votes, and regulatory development.
• Activity of the Congress, Senate, and regional parliaments: initiatives, session diaries, hearings, committees, as well as the role of different parliamentary groups.
• Public policies and sectoral regulation in Spain (for example, energy, digitalization, health, housing, transport, taxation, data protection, etc.), both from a normative perspective and from the political and media agenda.
• Spanish political and institutional actors: Government, ministries, senior officials, parties and their spokespersons, as well as their positioning in various public debates.
• Procedures and concepts of Spanish public law and the European Union, such as the functioning of the legislative procedure, the use of royal decree-laws, public consultations, hearings, and regulatory impact.

If you are interested in the issue of technological innovation and AI regulation from a comparative perspective, I can help you with the Spanish part, for example:

• Explain what powers the Government of Spain, the involved ministries (for example, Economy, Digital Transformation, Science, Justice) and independent authorities have in digital matters.
• Analyze how the application of the European AI Regulation (AI Act) is articulated in Spain, which bodies intervene, and what margin the Executive has to develop complementary regulations.
• Review Spanish regulatory projects or strategies on digitalization, data, cybersecurity, or artificial intelligence, explaining their content and relevance.

I can also help you to:

• Understand the role of the Government and the General Courts in regulating new technologies in Spain (who proposes, who approves, what controls exist).
• Prepare notes, summaries, or argumentaries on current debates in Spain related to AI and innovation (ethical risks, impact on employment, privacy, digital sovereignty, etc.).
• Identify which initiatives or policies affect a specific sector (for example, fintech, digital health, connected mobility) within the Spanish and European regulatory framework.

If you reformulate your question towards the Spanish or European Union scope — for example, asking about the powers of the Government of Spain or the European Commission in technological innovation and AI regulation, or about the Spanish laws that implement or complement the European framework — I will be able to give you a detailed and structured answer, adjusted to the format and length you need.

I cannot elaborate in detail on Sam Altman's professional and political career because my specialization and sources of work are focused on the institutional and political sphere of Spain (parliament, governments, parties, regulations, public actors and their activity), not on general biographies of figures from the international technology sector.

My information environment is designed to follow and analyze:

• Regulations and initiatives processed in Spain: draft and proposed bills, royal decree-laws, ministerial orders, public consultations, and government agreements.
• Parliamentary activity in the Congress, Senate, and regional parliaments: debates, hearings, questions, votes, and positions of the groups.
• Political and institutional actors: profiles of public officials, composition of committees and parliamentary groups, and their public activity related to specific policies.
• Political and regulatory news in Spain, with a special focus on how public decisions affect economic sectors, organizations, and citizens.

For that reason, I cannot guarantee you an accurate, updated, and sufficiently verified biography of Sam Altman or other personalities whose professional trajectory mainly belongs to the global technology and business field, and not to the Spanish political or institutional ecosystem.

However, I could help you with issues such as:

• Analyzing how artificial intelligence is entering the political and regulatory agenda in Spain or the European Union.
• Identifying which laws, strategies, or plans in Spain relate to AI, data, algorithms, digital rights protection, or technological innovation.
• Reviewing what certain parties, ministries, or senior officials in Spain have said about OpenAI, generative AI, or the regulation of large models.
• Preparing a political-regulatory summary for your organization about risks and opportunities opened by AI deployment in Spain.

If you are interested in moving the conversation towards that area, you can, for example, ask me:

• What regulatory framework exists in Spain and the EU on artificial intelligence and how it may affect companies using models like OpenAI.
• What position the main Spanish parties have expressed regarding AI and work automation.
• What role bodies like the European Commission or the Government of Spain play in defining rules that impact companies like OpenAI.

If you reformulate your question towards any of these angles (regulation, public policies, parliamentary debate, impact on sectors in Spain), I will be able to give you a much more complete and useful answer within my area of expertise.