Artificial intelligence is rapidly becoming part of everyday professional work. For law firms and consulting firms, the question is no longer simply whether to use AI, but how to use it effectively and responsibly within an increasingly demanding regulatory framework.
AI can analyze large volumes of information, summarize documents, assist with legal research, review contracts, structure arguments, detect patterns, and automate certain administrative tasks. Used well, it can allow professionals to spend less time on repetitive work and more time on activities where human judgment adds the most value: strategy, negotiation, interpretation, client relationships, and decision-making.
However, that same technology can create significant legal, reputational, and professional risks when its outputs are accepted without proper verification.
For this reason, the organizations likely to gain the greatest advantage from AI will not necessarily be those that adopt the most tools, but those that develop better governance, training, and professional judgment around its use.
Spain Is Already Showing Us What Happens When AI Is Used Poorly
A particularly telling example occurred at the High Court of Justice of the Canary Islands in February 2026.
An attorney filed an appeal containing 52 case law references. Of these, 48 turned out to be incorrect or nonexistent and had been obtained using a generative AI tool without the attorney properly verifying the results. The appeal also included a purported report from the General Council of the Judiciary that could not be located.
The court fined the attorney €420 for conduct contrary to the duty of truthfulness and the rules of procedural good faith, and referred the matter to the relevant bar association. EFE Noticias
The lesson is especially relevant: the problem was not using artificial intelligence. The problem was trusting it without checking its work.
Generative AI can produce extraordinarily convincing text that nevertheless contains false information. In the legal field, a fabricated court judgment, an incorrect decision number, or a nonexistent citation is not simply a minor technical error. It can weaken a legal argument, harm a client’s interests, and undermine the credibility of the professional who signs the document.
Spain had already received another significant warning in 2024. The First Chamber of the Constitutional Court issued a formal warning to an attorney who had included purported verbatim quotations from 19 of the Constitutional Court’s own judgments in a petition for constitutional protection. None of those quotations was real. The decision highlighted an essential issue: whatever the source of the error, responsibility for the content submitted to a court remains with the professional. (Constitutional Court)
The rule that emerges from both examples is clear:
AI can assist a professional, but it does not assume that professional’s responsibility.
There Are Also Examples of AI Being Used Well
Spain’s experience also offers positive examples.
The Judicial Documentation Center (CENDOJ) has developed KENDOJ—Knowledge Extractor for CENDOJ—a tool based on artificial intelligence techniques for processing legal documents in Spanish.
Among other functions, KENDOJ can structure court decisions, detect entities, identify links to case law and legislation, classify documents, and perform pseudonymization. The system combines natural language processing techniques, rules, and machine learning models. (Judiciary)
This type of application represents a very different model from the indiscriminate use of a generative tool.
Here, AI is used for a defined function within an institutional environment and with control procedures in place. In other words, there is a specific task, a clear purpose, and mechanisms for review.
This approach has been reinforced by Instruction 2/2026 of the General Council of the Judiciary on the use of artificial intelligence systems in judicial work.
The Instruction expressly establishes the principle of effective human oversight and stipulates that AI systems may not operate autonomously to make judicial decisions, assess facts or evidence, or interpret and apply the law.
At the same time, it allows certain authorized systems to be used as support tools to search for legal information, analyze and classify documents, generate outlines or summaries, and prepare certain internal drafts, always subject to human review. (BOE)
The idea, then, is not to prohibit AI.
It is to use it where it truly adds value and keep humans involved where judgment, responsibility, and decision-making are required.
Why AI Regulation Matters to Businesses
This issue is especially important for law firms and consulting firms because AI governance is becoming a matter of both competitiveness and regulatory compliance.
Article 4 of the European Artificial Intelligence Regulation—commonly known as the AI Act—sets out obligations relating to AI literacy.
Providers and deployers of AI systems must take steps to ensure that the people using those systems on their behalf have adequate knowledge of how they work, their capabilities, and their risks, taking into account their experience, training, and the specific context in which the systems are used. (Eur-Lex)
The AI literacy obligations began to apply on February 2, 2025, and the European framework emphasizes that organizations must train their employees with both the systems they use and the risks those systems may create in mind. (Digital Strategy)
For a professional services firm, this means that implementing artificial intelligence should involve much more than purchasing a software license.
A reasonable internal policy should specify, among other things:
- which AI tools may be used;
- what confidential information may or may not be entered into them;
- when human verification is mandatory;
- how legal references and data must be checked;
- who bears ultimate responsibility for the work produced;
- what training employees must receive;
- and what procedures must be followed for higher-risk applications.
The Prompt Can Also Contain Confidential Information
For law firms and consulting firms, there is another particularly sensitive issue: the information professionals enter into AI tools.
The Spanish Data Protection Agency has recently warned that even a simple prompt can contain personal data, confidential information, trade secrets, contracts, or documents subject to confidentiality obligations.
The AEPD itself points out that the emergence of generative AI is creating a new risk perimeter for organizations: the text box where an employee enters an instruction can also become an exit point for sensitive information. (AEPD Lab)
Asking an AI tool to “summarize this contract,” “analyze this claim,” or “review this case file” may seem harmless.
But depending on the tool used and the information entered, it can raise issues involving data protection, professional confidentiality, contractual confidentiality, or trade secrets.
For attorneys and consultants, the prompt must therefore begin to be treated as part of the company’s information security perimeter.
The Fears Are Real, but They Do Not Have to Define the Future
Concerns about artificial intelligence are understandable.
People fear that certain jobs will disappear, that professionals will become overly dependent on machines, that information will be leaked, that biased algorithms will play a role in important decisions, or that it will become increasingly difficult to distinguish authentic information from artificially generated content.
Some of these risks are significant and must be taken seriously.
But they should not lead us to assume that the future of AI will necessarily be negative.
The more interesting question is this: what kind of professional environment do we want to build around this technology?
Artificial intelligence is especially powerful at processing large amounts of information, detecting patterns, and producing initial analyses at great speed.
Humans, for their part, continue to bring essential capabilities: understanding context, questioning assumptions, applying ethical judgment, negotiating competing interests, building client relationships, interpreting ambiguous situations, and taking responsibility for decisions.
In the legal and consulting fields, combining these capabilities can be enormously powerful.
We can imagine an attorney spending less time manually reviewing hundreds of documents and more time developing the argument that will determine the outcome of a case.
Or a consultant who can quickly analyze an enormous amount of information and spend more time answering the question that really matters: what does all this mean for the client?
We can also imagine teams where younger professionals use AI to speed up certain routine tasks while learning to question, cross-check, and verify its outputs.
That future does not have to be framed as a competition between people and machines.
It can simply be a future in which professionals have better tools.
From AI Adoption to AI Maturity
The next stage of artificial intelligence in professional services should be less about novelty and more about organizational maturity.
A mature organization does not assume that everything an AI tool produces is correct.
But neither does it prohibit its use simply because it can make mistakes.
It identifies which tasks truly benefit from automation, selects the right tools, protects sensitive information, trains its employees, establishes verification mechanisms, and maintains meaningful human oversight of decisions that affect clients and individuals.
The competitive distinction may ultimately be relatively simple:
Using AI will become routine. Knowing when to trust it, when to question it, and when not to use it will be a professional skill.
For law firms and consulting groups, that skill will likely become increasingly important for protecting their clients’ trust, improving productivity, and continuing to provide high-quality professional advice.
Artificial intelligence does not need to replace human knowledge to transform a business.
Used well, it can amplify it.