1. Regulatory Framework in Spain

Spain is the most advanced EU country in AI governance:

A. EU AI Act — Directly Applicable
Regulation (EU) 2024/1689 is directly applicable in Spain without transposition. It classifies AI systems by risk and establishes 8 prohibitions related to subliminal techniques and vulnerability exploitation.

B. Spanish Enforcement Layer

• AESIA — Agencia Española de Supervisión de la Inteligencia Artificial: Created by Royal Decree 729/2023, headquartered in A Coruña, with sanctioning powers. It consolidates its role in Europe for ethical, sustainable and trustworthy AI.
• Draft Organic Law 26 May 2026: On 26 May 2026, Spain's Council of Ministers approved draft Organic Law on proper use and governance of AI, aligning national law with EU AI Act. AESIA will be single point of contact, with AEPD and CGPJ for biometrics, employment.
• MiFID II transposition: Royal Decree-Law 21/2017, art 9: Prevention of anomalies — algorithmic trading systems cannot create or contribute to disorderly trading conditions on the market.

C. Data and Labor Safeguards

• AEPD Guidance on Agentic AI (2025): If agentic AI introduces new recipients of personal data, changes retention periods, triggers new automated decisions or involves international transfers, individuals must be informed.
• Art. 64.4.d Workers' Statute + Ley 12/2021 Rider Law: Employer must inform works council of parameters, rules and instructions forming basis of algorithms or AI systems impacting working conditions. First law in Europe to regulate algorithmic management.

2. Ethics of Machine Economies — The Spanish Doctrine

Based on IIIA-CSIC Barcelona Declaration and CSIC AI Hub:

1. Opacity and Accountability: Technological opacity – difficulty understanding why a system makes a specific decision – is one of the main challenges when addressing liability arising from autonomous systems. Autonomous nature means they operate without uniform norms, balancing interests becomes probabilistic judgment.
2. Objectivation of Liability: For AI, Spain is moving to 'objective or quasi-objective liability' where mere risk of activity or benefit for owner is more important than actual guilt.
3. Human Centricity: AI systems must be driven by people's needs and values, and evolve with those evolving needs.
4. Fairness in Algorithmic Pricing: Dynamic pricing evolved from optimisation into core infrastructure — questions of fairness, transparency, consumer trust remain conceptually unresolved.

3. Case Law — Detailed

A) What is Money in a Machine Economy?

STS 326/2019, 20 June 2019 — First Bitcoin Judgment, Tribunal Supremo — Landmark.
Facts: Managing director received Bitcoin from 5 investors to reinvest via high frequency trading, failed to do so.

Held: 'Neither the so-called bitcoin is something susceptible to return, since it is not a material object, nor does it have the legal consideration of money'. Bitcoin is an intangible asset of consideration or exchange in any bilateral transaction that parties accept.

Ethical implication for Machine Economies: Machines cannot be paid in "money" under Spanish civil law if paid in crypto. Smart contracts paying machines in tokens are barter contracts, not payment. Restitution must be in euros equivalent at time of damage. This is why CNMV requires crypto ads to have license warnings.

B) Who Pays When the Machine Causes Damage?

Case 2: Tribunal Supremo — Liability for AI Malfunction in Autonomous Vehicles (2021)
Court held the manufacturer is liable under Product Liability Law, even without direct human negligence.

Supported by doctrine: In Spain there are not specific regulations regarding liability for damages or insurance for AVs, but general rules on insurance are applicable. Driver will be liable for decisions taken by autonomous vehicle, and manufacturer may also be liable.

And STS 816/2023? Actually jurisprudence on compensation: Compensation for damages shall follow general system of valuation, without injured party suffering loss depending on whether vehicle is autonomous.

Machine Economy Rule: In M2M crash — e.g., two delivery bots collide and order goods — Spain applies risk theory: owner / user of autonomous system + manufacturer are jointly liable under art 1905-1908 CC and Real Decreto Legislativo 8/2004 on motor vehicle insurance.

C) Are Smart Contracts Valid Between Machines?

General Rule — ILP Abogados / Spanish Civil Code
Spanish law does not prohibit smart contracts and recognises their validity in principle. Smart contracts must comply with existing regulations such as Spanish Civil Code.

Liability and Enforcement: Liability if smart contract does not perform as expected may depend on specific circumstances.

Case Law Gap: There is no Tribunal Supremo judgment yet that says "smart contract mortgage is valid". Lower courts treat code as expression of contract, but need human-readable terms under Ley 34/2002 LSSI and RD 1/2007 Consumers.

For Machine Economy: M2M self-executing contract is valid under art 1261 CC if consent, object, cause exist. Consent can be given by prior programming (pre-consent). Ethics: Need explainability — party must prove algorithm's parameters.

D) Algorithmic Collusion — When Machines Fix Prices

CNMC S/0637/18 (2018) — Uber / Cabify / Unauto
CNMC dismissed price-fixing allegations, finding their algorithms operated independently for dynamic pricing adjustments.

CNMC S/0003/20 (2021) — Real Estate Intermediaries
CNMC investigated intermediaries for using algorithmic platform to coordinate minimum fees and share sensitive information. First investigation in Spain where algorithm itself was tool for hub-and-spoke collusion.

Ethical principle established: Parallel use of same pricing algorithm is not per se collusion, but facilitating a common algorithm to coordinate commissions = anti-competitive under art 1 LDC and 101 TFEU. For Machine Economy: If two bots use same third-party repricing API, they can be liable for tacit collusion.

E) Algorithmic Trading — Disorderly Machines

RD-Law 21/2017 enforcement + CNMV 2024 Activity Plan
CNMV carried out analysis to detect use of AI in algorithmic trading in Spain — analysis of order books for more than a year.

Law: Investment firms must have kill-switch, throttling, prevent sending erroneous orders. Ethically, machine trading cannot exploit market structure.

Related: CNMV Fine X €5 Million — Quantum AI (2025)
CNMV fined platform X €5m for hosting ads from unlicensed crypto firm Quantum AI. Under Law 6/2023 on Securities Markets, platforms must scrutinize mass-reach financial ads. Important for machine economies because bots promote financial products — intermediary liability for AI-generated ads.

F) Agentic AI and Human Rights

Worldcoin — AEPD Order 6 March 2024 + Supreme Court Order 11 March 2024
AEPD prohibited Worldcoin data collection for 3 months. Worldcoin sought injunction before Contentious-Administrative Chamber. Supreme Court of Spain denied injunction, reiterating priority of "safeguarding of public interest".

Applied to Machine Economy: Machines collecting biometric data to transact (eye-scan to pay) requires prior Data Protection Impact Assessment. Machine-to-machine transfer of personal data is still processing — need Art 22 GDPR protection against solely automated decisions.

4. Ethical Checklist for Building a Machine Economy in Spain

If you are deploying in Spain, AESIA + AEPD will ask:

1. Provenance Label: Since 26 May 2026 draft, AI-generated content resembling real people/places must be labelled — serious infringement if not.
2. Liability Insurance: Objective liability — contract must allocate manufacturer / deployer / user chain. Cannot disclaim to end user.
3. No Disorderly Trading: Kill-switch, logs kept 5 years (MiFID).
4. Human Oversight: For high-risk (employment, credit scoring, biometric), human in the loop mandatory.
5. Algorithm Transparency to Workers: If your machine economy manages riders, drivers