The Spanish Legal Profession's ten principles for using AI, and how to apply them
In May 2026 the Advisory Council on Innovation and the Legal Profession of the Spanish Legal Profession published ten principles for the safe and responsible use of artificial intelligence. We summarise them and suggest how to turn them into concrete firm practices.
On 28 May 2026, the website of the Spanish Legal Profession published the presentation of a "Decalogue of principles for the safe and responsible use of AI systems", drawn up by its Advisory Council on Innovation and the Legal Profession (CAIA). The central idea is summed up by the news item itself: technology helps, but the one who answers remains the lawyer.
The decalogue is not a rule: it is a guide of principles. But it comes from the institution that represents the profession, and its ten points offer a useful framework for any firm that uses or is assessing artificial intelligence tools.
The ten principles
| No. | Principle | What it says, in brief |
|---|---|---|
| 1 | Effective human control | AI is a support tool, subject to real and effective supervision by the professional, and its results must be verified |
| 2 | Personal responsibility | The professional assumes full and exclusive responsibility for their actions; using AI does not relieve them of ethical, civil or criminal liability |
| 3 | Fundamental rights | Respect for effective judicial protection, equality, non-discrimination and data protection |
| 4 | Technical quality | Prioritise systems with certified and up-to-date sources, and know the origin of the data and the tool's limitations |
| 5 | Transparency | Proactively inform the client of the use of AI, indicating which tasks are assisted and which require direct human intervention |
| 6 | Explainability | Understand the basic workings of the systems in order to explain their results |
| 7 | Proportionality | The use must bring real value and the benefits must justify the risks |
| 8 | Confidentiality and security | Guarantee professional secrecy, data protection and cybersecurity |
| 9 | Bias prevention | Identify and avoid algorithmic biases that affect equality |
| 10 | Continuous training | Permanent updating on the capabilities, risks and ethics of AI |
From principles to practice
The principles are general. We suggest how to translate the most demanding into concrete firm decisions.
Human control and responsibility (principles 1 and 2)
No brief, report or email with AI-generated content should leave the firm without a lawyer having read it in full and checked every citation. Responsibility is personal: it is not shared with the tool or its provider.
Courts have already sanctioned briefs with non-existent citations, as we covered in the review of sanctions for false case law. And the Legal Profession itself has specified the ethical scope of this failure to verify in a circular we analyse in the news item on the ethics circular on AI.
Transparency with the client (principle 5)
The decalogue asks for proactive information to the client. A natural place to do it is the engagement letter: a clause explaining which tools are used, for which tasks and with what confidentiality guarantees.
Confidentiality (principle 8)
This is the principle that demands the most operational decisions: which tools are authorised, what data may be entered into each and under what contractual conditions. The Spanish Data Protection Agency's guidance on generative AI, which we summarise in the article on generative AI and data protection, is a good complement.
Training (principle 10)
Training is not only technical. It includes knowing when not to use the tool: in matters the lawyer does not master, AI multiplies the risk of failing to detect an error.
An outline for an internal policy
With the decalogue as a base, a firm's AI policy could have five sections:
- Authorised tools and permitted uses of each.
- Data never entered into external tools.
- Mandatory review of every result before use.
- Information to the client in the engagement letter.
- Training and periodic review of the policy.
We argued for the need for this policy in the editorial on the AI policy at the firm. More analysis in Legal AI.
Fuentes
- Consejo General de la Abogacía Española. Decálogo de principios para el uso seguro y responsable de sistemas de IA (PDF). Accessed 18 October 2026.
- Abogacía Española. La Abogacía Española lanza un decálogo sobre el uso ético de la IA en la práctica jurídica (28 de mayo de 2026). Accessed 18 October 2026.
The specific data in this article come from the official sources cited.
