LEGAL ARTIFICIAL INTELLIGENCE

A litigation practice built on decades of craft and on technology of our own.

Our partners have acted in some of the most complex proceedings of recent decades, and have learned from them that the advantage in a large-scale case is often decided by the ability to master the information. Constitución 23 has therefore built its own technological environment, combining artificial intelligence models, direct access to the official legal sources and verification procedures, at the service of its lawyers’ judgment.

A conventional search engine returns words. Our system understands the legal question, checks it against the official sources and returns the decisions and rules that answer it, with their full text and their official link. Among other sources, it consults:

• The case law of the Spanish judiciary, with judgments of the Supreme Court, the National High Court, the High Courts of Justice of the autonomous regions and the Provincial Courts.
• The Constitutional Court, the Court of Justice of the European Union and the European Court of Human Rights.
• The Official State Gazette and consolidated legislation, comparing the successive wordings of each rule and checking whether it is still in force, together with the regional gazettes.
• The Directorate-General for Taxation, the Economic-Administrative Courts and the Public Procurement Appeal Tribunals, the opinions and circulars of the State Prosecution Service, the opinions of the Council of State and land registry information.

The system does not merely locate decisions. It reconstructs how the case law on a question has evolved, identifies which judgments cite it and which have departed from it, and calculates procedural time limits, interest and sentences under the rules in force at each moment.

An investigation running to thousands of pages, years of a company’s email, dozens of hours of recordings or a complete set of accounts are materials that no lawyer can master alone within the time available. Our system processes them as follows:

• Optical character recognition of every document, with quality control of the reading.
• A complete index of the file and search by wording and by meaning, with a reference to the exact page.
• Transcription of hearings, statements and recordings, identifying each speaker and with time stamps.
• A direct link from any pleading to the page of the file that supports each assertion.
• Preparation of examinations with the question, the document and the exact page to be shown at each moment.

The result is that the team knows the whole file, not a sample of it, and that every assertion can be checked in seconds.

Artificial intelligence can get things wrong, and the courts know it. We have therefore set a rule that admits no exceptions: nothing is asserted in a pleading that has not been checked against the original source. Every citation of case law is checked against the official decision, every quoted passage against its literal wording and every reference to the file against the page of the record. The result is work that is faster and, above all, more reliable.

Professional secrecy admits of no technological exceptions. The analysis of case files and the transcription of recordings run on infrastructure under the firm’s control, access is restricted, and any third-party tool is used with contractual safeguards and data minimisation. We comply with the General Data Protection Regulation and the European Artificial Intelligence Act, and every piece of work assisted by these tools is supervised and signed by a lawyer.