Hybrid Human-Ai Advisory Liability Allocation Spain Banking Law .
HYBRID HUMAN–AI ADVISORY LIABILITY ALLOCATION UNDER SPAIN BANKING LAW
1. Introduction
Hybrid Human–AI Advisory Liability Allocation refers to the legal framework governing responsibility when financial institutions use artificial intelligence systems together with human financial advisers to provide banking, investment, credit, wealth-management and financial-product recommendations.
In Spain, banks increasingly use AI-based technologies for:
Investment portfolio recommendations.
Automated financial planning.
Creditworthiness and affordability assessments.
Mortgage and consumer-credit recommendations.
Algorithmic risk profiling.
Robo-advisory and wealth management.
Automated fraud and transaction-risk assessments.
Personalised financial-product recommendations.
Hybrid advisory arrangements involve two principal decision-makers:
AI systems that analyse financial information, calculate risks and generate recommendations.
Human advisers who review, approve, modify, communicate or implement those recommendations.
The central legal question is:
When an AI system generates an unsuitable financial recommendation and a human adviser approves or communicates it, who bears legal responsibility: the bank, the human adviser, the AI provider or the technology developer?
Spanish law does not recognise AI as an independent legal person capable of bearing ordinary civil liability. Consequently, responsibility generally remains with the financial institution, its contractual counterparties and other legally responsible persons, depending on their respective conduct and obligations.
The relevant framework combines Spanish banking law, investment-services regulation, civil liability, consumer protection, data protection and European Union AI regulation.
2. Legal Framework Governing Hybrid AI Advisory in Spain
A. Law 6/2023 on Securities Markets and Investment Services
The Spanish Securities Markets and Investment Services Law (Ley 6/2023) is a central legislative instrument governing investment services.
It implements the relevant EU investment-services framework and establishes obligations for investment firms and other regulated entities.
Its relevance to hybrid AI advisory includes:
Investor protection.
Suitability and appropriateness assessments.
Information and disclosure obligations.
Conflicts-of-interest management.
Organisational requirements.
Recordkeeping and supervisory controls.
Where a Spanish bank provides investment advice using AI-generated recommendations, it cannot avoid its regulatory obligations merely because the recommendation originated from an algorithm.
The institution must ensure that the recommendation complies with applicable investor-protection requirements.
B. MiFID II – Directive 2014/65/EU
MiFID II is particularly important for determining responsibility in AI-assisted investment advice.
Article 24 establishes the general obligation to act honestly, fairly and professionally in accordance with the best interests of clients.
Article 25 establishes requirements concerning:
Knowledge and competence.
Suitability assessments.
Appropriateness assessments.
Client information.
Investment recommendations.
For hybrid AI advisory, these obligations mean that a bank must ensure that its automated recommendation process produces advice consistent with the client's investment objectives, financial circumstances, risk tolerance and knowledge.
A human adviser cannot simply approve an AI-generated recommendation without adequate verification.
C. Commission Delegated Regulation (EU) 2017/565
This Regulation supplements MiFID II and establishes detailed organisational and conduct requirements.
Its provisions are relevant to:
Client profiling.
Investment recommendations.
Suitability assessments.
Compliance arrangements.
Internal controls.
Recordkeeping.
Conflicts of interest.
A financial institution using AI must maintain adequate procedures to establish that its advisory systems operate consistently with these obligations.
D. Spanish Civil Code
The Spanish Civil Code provides the foundation for civil liability arising from defective financial advice.
Important provisions include:
Article 1101: Establishes contractual liability for persons who, in fulfilling their obligations, act fraudulently, negligently or in breach of their contractual obligations.
Article 1104: Defines negligence by reference to the degree of diligence required by the nature of the obligation and the circumstances of the persons, time and place.
Article 1258: Requires contracts to be performed in accordance with good faith and includes consequences arising from the nature of contractual obligations.
Article 1902: Establishes non-contractual liability for damage caused through an act or omission involving fault or negligence.
These provisions are particularly important where AI-generated financial advice causes investment losses or other financial damage.
E. General Data Protection Regulation (GDPR)
Regulation (EU) 2016/679 applies where AI advisory systems process personal data.
Relevant provisions include:
Article 5: Principles governing personal-data processing.
Article 6: Lawfulness of processing.
Article 13: Information obligations.
Article 14: Information where personal data are not obtained directly from the individual.
Article 15: Access rights.
Article 22: Automated individual decision-making, including profiling.
Article 35: Data protection impact assessments.
Article 82: Right to compensation for damage.
Article 22 is particularly relevant where a decision based solely on automated processing produces legal effects or similarly significantly affects an individual.
The provision is subject to specified exceptions and safeguards. Not every AI-assisted recommendation automatically falls within Article 22.
F. EU Artificial Intelligence Act
Regulation (EU) 2024/1689 establishes the European framework for AI systems.
Its relevance to banking includes:
AI governance.
Risk classification.
Transparency.
Human oversight.
Technical documentation.
Monitoring and risk management.
Responsibilities of providers and deployers.
Certain AI systems used to evaluate the creditworthiness of natural persons or establish their credit scores are classified as high-risk systems, subject to the applicable statutory exceptions.
However, an AI system used exclusively to provide investment recommendations does not automatically become high-risk merely because it operates in a bank.
The classification depends on the system's intended purpose and the relevant provisions of the Regulation.
Banks acting as deployers must comply with the obligations applicable to their role. AI providers have separate obligations.
G. Digital Operational Resilience Act (DORA)
Regulation (EU) 2022/2554 applies to financial entities, including relevant Spanish banking institutions.
DORA is important because hybrid advisory depends on digital infrastructure, external technology providers and information and communication technology systems.
Its principal requirements concern:
ICT risk management.
Incident reporting.
Operational resilience.
Third-party ICT risk.
Business continuity.
Digital-system testing.
Governance and accountability.
A bank cannot automatically transfer its regulatory responsibility to an external AI technology provider through an outsourcing contract.
3. Meaning of Hybrid Human–AI Advisory Liability Allocation
Hybrid advisory liability allocation means identifying which parties are legally responsible when an AI-assisted financial recommendation causes loss, regulatory non-compliance or consumer harm.
The principal parties are:
1. Banking institution
The bank is generally responsible for ensuring that its advisory services comply with applicable banking and investment-services regulations.
2. Human financial adviser
The adviser may incur responsibility where professional negligence, misleading communication, inadequate verification or breach of professional duties is established.
3. AI system provider
The technology provider may incur contractual or other applicable legal liability where defective software, inadequate documentation, security failures or other actionable failures cause damage.
4. AI system deployer
The deployer is generally the financial institution that uses the AI system in its business operations. It has obligations under the AI Act where applicable.
5. Data and technology suppliers
Third-party providers may become liable where defective data, inadequate services or contractual breaches contribute to the loss.
These parties do not necessarily share liability equally. Their responsibility depends on the relevant legal duty, causal contribution, contractual relationship and applicable liability rules.
4. Allocation of Liability Between Banks and Human Advisers
A. Responsibility of the Banking Institution
The banking institution generally has primary responsibility towards its customers for the financial advisory services it provides.
For example:
A Spanish bank uses an AI platform to analyse a customer's financial position and recommend a high-risk investment product.
The AI system incorrectly assesses the customer's risk tolerance and recommends an unsuitable investment.
A bank employee approves the recommendation and presents it to the customer as suitable.
If the customer suffers financial losses, the bank may face liability based on:
Breach of investment-services obligations.
Failure to conduct an adequate suitability assessment.
Negligent advisory services.
Inadequate internal supervision.
Misleading or incomplete information.
Breach of contractual obligations.
The bank cannot ordinarily defend itself merely by arguing that the recommendation was generated by an external AI system.
B. Responsibility of Human Financial Advisers
Human advisers remain subject to applicable professional and contractual duties.
Their potential liability depends on their actual role.
For example, an adviser who identifies an obvious inconsistency in an AI-generated recommendation but approves it without investigation may have contributed to negligent advice.
Relevant considerations include:
Whether the adviser had access to the customer's actual financial information.
Whether the adviser understood the product being recommended.
Whether the AI recommendation contained identifiable errors.
Whether the adviser had authority to reject the recommendation.
Whether the adviser independently assessed suitability.
Whether the adviser accurately explained the risks.
However, a human adviser should not automatically be held liable merely because an AI system made an incorrect calculation.
The adviser's responsibility must be assessed in light of their legal duties, authority, conduct and causal contribution.
C. Responsibility of the AI Technology Provider
An AI provider may face liability where its conduct or defective product causes legally recognised damage.
Potential grounds include:
Defective algorithmic design.
Material software defects.
Incorrect technical implementation.
Failure to provide contractual functionality.
Inadequate security safeguards.
Misrepresentation concerning system capabilities.
Failure to comply with applicable product or AI regulatory requirements.
The bank may also pursue contractual remedies against the provider.
The provider's liability towards the customer is not necessarily identical to its liability towards the bank. The contractual structure and applicable product-liability and civil-liability rules must be examined separately.
5. Suitability and Appropriateness in AI-Assisted Investment Advice
Suitability is one of the most important legal issues in hybrid advisory.
Under MiFID II, investment firms providing investment advice or portfolio management must obtain the information necessary to assess whether the recommendation is suitable.
The relevant considerations include:
Investment knowledge.
Investment experience.
Financial situation.
Ability to bear losses.
Investment objectives.
Risk tolerance.
Example
A customer has limited investment experience and requires capital preservation.
The bank's AI system recommends a complex leveraged financial instrument because its algorithm predicts potentially high returns.
The human adviser approves the recommendation without properly considering the customer's financial circumstances.
This may constitute a breach of suitability obligations.
The bank cannot establish compliance merely by showing that the AI system generated a recommendation based on statistical calculations.
The institution must demonstrate that the legally required assessment was properly conducted.
Appropriateness Assessment
Appropriateness generally concerns whether a client has the necessary knowledge and experience to understand the risks of a particular investment product or service where the relevant assessment is required.
An AI system may assist in evaluating customer responses, but the financial institution remains responsible for applying the applicable legal standards.
6. Human Oversight and the Meaning of Meaningful Human Intervention
Human oversight is a central element of responsible AI deployment.
A human adviser must have an effective opportunity to assess and, where necessary, reject or modify an AI recommendation.
Merely placing a human employee at the end of an automated process does not necessarily constitute meaningful oversight.
Effective oversight should involve:
Access to relevant customer information.
Understanding of the AI system's purpose and limitations.
Ability to identify potentially incorrect outputs.
Authority to reject unsuitable recommendations.
Access to escalation procedures.
Adequate training.
Proper documentation of human intervention.
The AI Act contains human-oversight requirements for relevant high-risk AI systems.
The precise obligations depend on the system's classification and the role of the bank.
A bank must distinguish between genuine human review and purely formal approval of automated recommendations.
7. Liability for Algorithmic Bias and Discriminatory Financial Advice
AI advisory systems may reproduce discriminatory patterns present in their training data.
For example:
A lending algorithm uses geographical information, employment history and other variables to estimate creditworthiness.
Although the system does not directly use a protected characteristic, certain variables may operate as proxies for protected characteristics.
This may result in indirect discrimination.
Relevant legal instruments include:
Spanish Constitution, Article 14.
Law 15/2022 on Equal Treatment and Non-Discrimination.
GDPR.
EU equality legislation.
Applicable AI Act requirements.
A financial institution must ensure that automated advisory and credit-assessment systems do not unlawfully discriminate against customers.
Where discrimination is established, the bank may face regulatory, civil or other legal consequences depending on the circumstances.
8. Contractual Allocation of Responsibility Between Banks and AI Providers
Spanish financial institutions frequently use external technology providers.
Contracts between banks and AI providers should clearly establish responsibility for:
A. Algorithmic performance
The provider should specify the system's capabilities, limitations and expected performance.
B. Data accuracy
The contract should identify responsibilities for data quality, validation and correction.
C. Regulatory compliance
The parties should allocate responsibility for compliance with applicable AI, data-protection and banking regulations.
D. Audit rights
Banks should retain adequate rights to examine relevant systems, documentation and controls.
E. Incident reporting
The provider should notify the bank of significant failures, security incidents and material system changes.
F. Indemnification
The agreement may establish compensation mechanisms for losses caused by contractual breaches.
G. Termination rights
Banks should have appropriate rights to suspend or terminate systems that create unacceptable legal or operational risks.
Importantly, contractual allocation between the bank and its provider does not automatically eliminate regulatory obligations owed by the bank to customers or supervisory authorities.
9. Regulatory Supervision in Spain
Several institutions play important roles in supervising hybrid AI advisory activities.
A. Banco de España
The Banco de España supervises credit institutions within its statutory responsibilities and the relevant European supervisory framework.
Its responsibilities include banking supervision, financial stability and applicable conduct-related supervisory functions.
It may examine governance, internal controls, risk management and the use of technology in banking operations.
B. Comisión Nacional del Mercado de Valores (CNMV)
The CNMV is particularly important where AI systems are used in investment services.
Its supervisory interests include:
Investor protection.
Investment-advice compliance.
Suitability procedures.
Conflicts of interest.
Market conduct.
Organisational requirements.
C. Agencia Española de Protección de Datos (AEPD)
The AEPD supervises compliance with data-protection legislation in Spain.
Its responsibilities become relevant when AI advisory systems process personal data, conduct profiling or make decisions falling within GDPR requirements.
D. European Banking Authority (EBA)
The EBA develops regulatory standards and supervisory guidance relevant to digital banking, operational risk, outsourcing and governance.
Its work is important for promoting consistent supervisory expectations across EU banking institutions.
10. Important Case Laws and Judicial Principles
There is an important distinction: the following judgments establish principles concerning financial advice, investment-services duties, contractual responsibility and automated decision-making. They should not be misrepresented as judgments directly deciding liability for modern hybrid human–AI banking advice.
Case 1: Genil 48 SL and Others v Bankinter SA and Others
Case C-604/11, Court of Justice of the European Union, judgment of 30 May 2013.
This is an important authority concerning investment-services obligations under MiFID.
The dispute concerned financial instruments and the obligations of investment firms towards customers.
The Court examined the regulatory requirements relating to suitability and appropriateness assessments.
Legal principles:
Investment firms must comply with applicable client-protection requirements.
Suitability and appropriateness obligations serve different purposes.
The applicable assessment depends on the nature of the investment service.
National courts must determine the consequences of regulatory breaches under the applicable national legal framework.
Relevance to AI advisory:
A bank cannot avoid MiFID obligations simply because an AI platform generates its investment recommendations.
If an AI-assisted recommendation fails to comply with suitability requirements, the bank's regulatory responsibilities remain relevant.
Case 2: Banif Plus Bank Zrt v Csipai and Csipai
Case C-472/11, CJEU, judgment of 21 February 2013.
The judgment concerned consumer protection and the judicial treatment of unfair contractual terms.
Legal principles:
The Court emphasised the importance of effective consumer protection and the role of national courts in addressing unfair contractual terms.
Relevance to AI advisory:
A financial institution cannot necessarily rely on standard contractual clauses to exclude responsibilities arising from unlawful or unfair advisory practices.
An AI-service agreement or customer disclaimer must be assessed under applicable consumer-protection legislation.
Case 3: Andriciuc and Others v Banca Românească SA
Case C-186/16, CJEU, judgment of 20 September 2017.
The dispute concerned foreign-currency loan agreements and transparency requirements.
Legal principles:
The Court examined whether consumers received sufficient information to understand the economic consequences and risks of contractual terms.
It emphasised that transparency requires consumers to be able to understand the potential economic consequences of contractual arrangements.
Relevance to AI advisory:
Where a bank uses AI to recommend foreign-currency loans, variable-rate products or complex financial instruments, customers must receive sufficiently clear information about material risks.
An AI-generated explanation does not automatically satisfy the bank's transparency obligations.
Case 4: Hirmann v Immofinanz AG
Case C-174/12, CJEU, judgment of 19 December 2013.
The case concerned investor protection and the legal consequences of misleading information in the context of company capital.
Legal principles:
The Court considered the relationship between investor-protection rules and national remedies.
Relevance to AI advisory:
Financial institutions should ensure that automated investment communications do not contain materially misleading information.
Where incorrect AI-generated information is communicated to investors, applicable investor-protection and national liability rules must be considered.
Case 5: Banco Español de Crédito SA v Joaquín Calderón Camino
Case C-618/10, CJEU, judgment of 14 June 2012.
The case concerned consumer protection and unfair contractual terms.
Legal principles:
The Court emphasised the importance of effective judicial protection against unfair terms in consumer contracts.
Relevance to AI advisory:
A bank should not assume that contractual clauses transferring all responsibility for automated recommendations to customers will necessarily be enforceable.
The validity of such clauses must be assessed under EU and Spanish consumer-protection rules.
Case 6: Spanish Supreme Court – Financial Derivatives and MiFID Advisory Obligations
Supreme Court judgment of 20 January 2014, STS 840/2013 (ROJ STS 354/2014).
This judgment belongs to the Spanish Supreme Court's significant line of decisions concerning financial derivatives and the information duties of financial institutions.
The relevant judicial approach concerns the importance of adequate information and the investor's ability to understand the risks of complex financial products.
Legal principles:
Financial institutions must comply with applicable information obligations.
The customer's knowledge and experience are relevant to assessing the quality of the advisory relationship.
Inadequate information may have consequences for the validity of consent and contractual liability, depending on the circumstances.
Relevance to AI advisory:
If a bank's AI system generates an explanation that materially understates the risks of a derivative and the adviser communicates it without proper verification, the bank may face legal consequences.
The case does not directly decide AI liability but provides relevant principles for assessing defective financial advice.
Case 7: Spanish Supreme Court – Financial Advice and Investor Consent
STS 491/2015, 15 September 2015.
This judgment forms part of the Spanish Supreme Court's jurisprudence concerning investment products and the information required for informed investor consent.
Legal relevance:
Financial-product risks must be adequately communicated, and the legal assessment cannot be reduced to the customer's signature on contractual documentation.
Application to AI:
A bank cannot rely exclusively on an automated disclosure screen or an AI-generated risk statement to demonstrate that a customer genuinely received the legally required information.
The quality, completeness and intelligibility of the information remain important.
11. Liability for Incorrect AI Predictions
A separate issue concerns financial predictions generated by AI.
For example:
An AI advisory platform predicts that a particular investment portfolio will produce an annual return of 12%.
The customer invests based on the recommendation, but the portfolio loses 25%.
A financial loss does not automatically establish legal liability.
The court must consider:
Whether the recommendation was represented as a guarantee or an estimate.
Whether the bank complied with its advisory obligations.
Whether the AI model was reasonably designed and tested.
Whether the customer received adequate risk disclosures.
Whether the adviser independently reviewed the recommendation.
Whether a breach caused the claimed damage.
Whether the loss resulted from ordinary market risk or negligent advice.
Banks are not generally insurers against investment losses.
However, they may be responsible where defective advisory processes or breaches of applicable duties cause compensable damage.
12. GDPR Liability for Automated Financial Recommendations
AI-based advisory systems frequently process personal and financial information.
The GDPR imposes important obligations on banks.
Article 5 – Data Processing Principles
Financial institutions must ensure that personal data are processed lawfully, fairly and transparently.
Data must be accurate and adequate for the relevant purpose.
Article 22 – Automated Decision-Making
Article 22 provides protections against certain decisions based solely on automated processing, including profiling, where the decision produces legal effects or similarly significantly affects the individual.
For example, an automated credit decision that significantly affects a customer's ability to obtain financial services may raise Article 22 issues.
A recommendation merely generated by AI and subsequently subject to genuine human assessment does not automatically constitute a decision based solely on automated processing.
Article 82 – Compensation
Article 82 provides a right to compensation for material or non-material damage caused by an infringement of the GDPR, subject to its requirements.
A bank may therefore face data-protection liability where unlawful processing or defective automated decision-making causes compensable damage.
13. AI Act and Banking Advisory Accountability
The EU AI Act introduces an additional regulatory dimension.
Its classification framework distinguishes between prohibited AI practices, high-risk AI systems, systems subject to transparency requirements and other systems.
High-Risk Creditworthiness Assessment
AI systems intended to evaluate the creditworthiness of natural persons or establish their credit scores are generally included in Annex III, subject to the statutory exception for AI systems used for detecting financial fraud.
This is particularly relevant to Spanish banks.
Banks deploying such systems must comply with the applicable high-risk AI requirements.
These include, as applicable:
Risk-management systems.
Data governance.
Technical documentation.
Logging.
Transparency and information.
Human oversight.
Accuracy and cybersecurity.
Post-market monitoring.
Investment Advisory Systems
Investment advisory systems must be classified according to their intended purpose and the specific provisions of the AI Act.
An investment recommendation system is not automatically high-risk merely because it is used by a bank.
Nevertheless, other financial-services obligations, including MiFID II and GDPR, continue to apply.
14. DORA and Third-Party AI Provider Responsibility
DORA is particularly significant where a Spanish bank obtains AI services from an external technology company.
The bank must manage risks arising from external ICT service providers.
Important areas include:
ICT third-party risk management.
Contractual requirements.
Access and audit rights.
Incident management.
Business continuity.
Exit strategies.
Concentration risk.
Operational resilience.
For example, if an external AI provider experiences a technical failure that produces incorrect customer recommendations, the bank must assess and manage the resulting operational and regulatory risks.
The bank cannot assume that outsourcing removes its supervisory accountability.
15. Practical Liability Allocation Matrix
| Situation | Potentially responsible party |
|---|---|
| AI produces an unsuitable investment recommendation | Bank; potentially AI provider |
| Human adviser approves an obviously defective recommendation | Bank and potentially adviser |
| AI provider supplies defective software | Provider, subject to applicable liability requirements |
| Bank fails to supervise its AI advisory system | Bank |
| Customer information is processed unlawfully | Relevant GDPR controller or processor, depending on its obligations |
| Algorithm produces discriminatory lending outcomes | Bank and potentially other responsible actors |
| AI system fails because of inadequate cybersecurity | Bank and potentially ICT provider |
| Adviser misrepresents investment risks | Bank and potentially adviser |
| External provider breaches its service contract | AI provider under the applicable contract |
| Bank fails to provide required human oversight | Bank and potentially other responsible actors |
The table identifies potential responsibility, not automatic liability. Each claim requires examination of the relevant legal requirements.
16. Defences Available to Financial Institutions
Spanish banks may raise several legal arguments when defending claims involving AI-assisted advice.
A. Absence of Causation
The bank may argue that the customer's loss resulted from market movements rather than defective advice.
B. Adequate Suitability Assessment
The bank may demonstrate that the customer's investment objectives, knowledge, financial situation and risk tolerance were properly assessed.
C. Adequate Risk Disclosure
The bank may demonstrate that material risks were clearly communicated before the transaction.
D. Independent Customer Decision
The bank may argue that the customer independently decided to undertake the investment after receiving adequate information.
However, customer choice does not automatically eliminate liability for a bank's own regulatory or contractual breaches.
E. Compliance with Internal Controls
The bank may provide evidence of model validation, human review, monitoring and compliance procedures.
F. Third-Party Responsibility
The bank may pursue a contractual claim against its AI provider where the provider's breach contributed to the loss.
However, this does not automatically defeat the customer's claim against the bank.
17. Compliance Requirements for Spanish Banks
A robust hybrid AI advisory compliance programme should include:
1. AI governance policy
Every bank should establish clear governance arrangements for the use of AI in financial advisory.
2. Human oversight procedures
Human advisers must have sufficient authority, competence and information to review AI-generated recommendations.
3. Suitability verification
AI recommendations must be subject to the applicable MiFID suitability requirements.
4. Algorithmic testing
Banks should conduct appropriate validation and monitoring to identify inaccurate or discriminatory outputs.
5. Documentation
The institution should retain appropriate records of recommendations, customer profiles, human interventions and relevant decisions.
6. Data protection
Personal information must be processed in accordance with the GDPR.
7. Third-party risk management
AI outsourcing arrangements must comply with relevant DORA requirements.
8. Staff training
Financial advisers must understand AI limitations and their continuing professional responsibilities.
9. Customer transparency
Customers should receive clear information about the nature and limitations of AI-assisted advisory where required by applicable law.
10. Complaint-handling procedures
Banks should provide effective procedures for investigating customer complaints involving automated recommendations.
18. Important Legal Distinction: Regulatory Liability and Civil Liability
A major issue in hybrid advisory disputes is the difference between regulatory liability and civil liability.
Regulatory Liability
Regulatory liability may arise where a bank breaches supervisory, organisational, conduct or AI-related requirements.
Depending on the infringement, supervisory authorities may impose administrative measures or sanctions.
Civil Liability
Civil liability concerns compensation for damage suffered by the customer.
A customer may rely on contractual or non-contractual liability principles, subject to the applicable legal requirements.
Data Protection Liability
GDPR liability may arise independently where unlawful personal-data processing causes compensable damage.
Contractual Liability Between Bank and Provider
A bank may also bring a separate contractual claim against its AI technology provider.
These forms of liability may coexist, but a regulatory breach does not automatically establish every element of a customer's damages claim.
19. Major Legal Challenges in Spain
The development of hybrid human–AI advisory creates several important legal challenges.
First: Attribution of responsibility.
It may be difficult to identify whether an incorrect recommendation resulted from defective programming, inaccurate data, inadequate human review or poor institutional governance.
Second: Explainability.
Complex AI systems may produce recommendations that are difficult for customers and advisers to understand.
Third: Human oversight.
There is a risk that employees simply approve algorithmic recommendations without meaningful independent assessment.
Fourth: Algorithmic discrimination.
Historical financial data may reproduce discriminatory patterns.
Fifth: Evidence and recordkeeping.
Customers may find it difficult to obtain evidence concerning the internal operation of an AI system.
Sixth: Outsourcing accountability.
Banks may depend heavily on technology providers, creating uncertainty over contractual responsibility and operational control.
Seventh: Cross-border supervision.
AI providers, banks and data-processing infrastructure may be located in different EU Member States or third countries.
Eighth: Compensation.
Determining whether an investment loss resulted from defective advice or ordinary market risk can be legally and factually complex.
20. Conclusion
Hybrid Human–AI Advisory Liability Allocation under Spanish Banking Law is governed by a combination of national civil law, EU investment-services legislation, data-protection rules, AI regulation and digital operational resilience requirements.
The central legal principle is that the use of artificial intelligence does not automatically remove the responsibility of a regulated financial institution.
Spanish banks remain responsible for complying with applicable investment-advisory, consumer-protection and organisational requirements.
Human advisers may also face responsibility where their own conduct breaches applicable professional or contractual duties.
AI technology providers may incur separate contractual or other legal liability where their products or services cause legally recognised damage.
The most important legal authorities include Law 6/2023, MiFID II, the GDPR, the EU AI Act, DORA and Articles 1101, 1104, 1258 and 1902 of the Spanish Civil Code.
The decisions in Genil 48, Andriciuc, Hirmann and the Spanish Supreme Court's financial-derivatives jurisprudence provide important principles for assessing investor protection, suitability, transparency and financial-advisory liability.
Ultimately, Spanish banking law does not permit a financial institution to use AI as a general defence against responsibility for defective financial advice. Liability must be determined by examining the relevant legal duties, the role of each participant, the existence of a breach and the causal connection with the damage suffered.

comments