Legal Business Blogs

Sponsored briefing: Artificial intelligence (AI) in civil proceedings – a forecast

Heuking Kühn Lüer Wojtek gives an overview on artificial intelligence in civil proceedings and discusses the current obstacles of using AI within the law

The status quo

For many years, the legal sector has felt safe from technical innovations that have revolutionised other sectors in the last three decades, making countless activities redundant. Technical innovations in the legal sector over the last few decades served to speed up and facilitate work processes. E-mail led to enormous acceleration of written communication, and electronic lawyers’ mailboxes replaced the physical submission of documents. None of these innovations fundamentally changed the actual work of lawyers: determining the facts with subsequent application of the law has long remained unaffected. This may soon change.

While using databases for legal research or search engines to check contracts for specific clauses (continues to) fall into the category of facilitating work, the application of the law itself remains firmly in lawyers’ hands. The same applies to mass litigation with thousands of similar cases where all parties are operating with text modules. The lawyer determines the facts of the case and combines the pre-set text modules.

Requirements for AI in litigation

The use of AI requires that the parties provide all pleadings and evidence in machine-readable form and, to the extent possible, in a uniform file format. A working group of presidents of German Higher Regional Courts and the Federal High Court of Justice already recommended that the parties present their facts and legal arguments in one preformatted document as basis for the proceedings. The parties shall then work simultaneously in the same document and each put their objections on law or fact at the right place. The use of such a uniform document should then allow the software to structure, combine and evaluate party submissions.

Once the software has extracted and compared all relevant text passages, it can check whether a particular element of the law is applicable. The judge may then concentrate on his individual legal assessment without having to spend time on structuring the file. While the use of such software will considerably accelerate the handling of cases in courts, the parties could also take advantage of such software. Lawyers can estimate their chances of success in court. This will help their clients to make an informed decision if the legal costs and efforts are worthwhile. The frequently observed rational lack of interest in the case of expected low yields is counteracted, and access to the law is facilitated.

Current obstacles to the use of AI

Even intelligent software may fail where complex issues are at stake that do not fall into simple yes/no categories. Software can therefore only be used once the facts have been established with sufficient probability.

Hurdles a software currently cannot overcome also exist within the law itself. There are not only complex statutory provisions that remain beyond the comprehension of AI, but also provisions that are open to interpretation and leave discretionary powers to the judge. It seems unlikely that AI will be able to penetrate these structures in the foreseeable future. Law that contains many descriptive or normative elements with simple definitions is better suited to AI than complex statutes that cannot be decrypted without prior knowledge.

Law that contains many descriptive or normative elements with simple definitions is better suited to AI than complex statutes that cannot be decrypted without prior knowledge.

An additional factor that makes it harder to predict the outcome of cases in Germany compared to common law jurisdictions is that lower courts do not have to follow precedents by other (higher) courts (although they usually voluntarily do). In the Volkswagen emissions scandal, for example, courts came to entirely different legal conclusions despite almost identical facts. That only changed after a Federal High Court of Justice landmark decision.

There are further serious constitutional objections to the use of AI in the judiciary. Article 101 (1) Basic Law grants access to a court of law. It is common opinion that this has to be a human. Some commentators view judicial procedural sovereignty and independence under article 97 Basic Law at risk. Because of judicial independence, it will likely also be impermissible to require judges to use AI. Additionally, the right to legal protection under article 20 (3) Basic Law in conjunction with article 2 Basic Law is only guaranteed if every citizen has access to the digital portals and can upload documents.

The constitutional problem becomes particularly acute when judges can no longer justify their decisions independently but refer to the allegedly superior AI. The problem of non-transparency of automated systems also arises in cases of automated court decisions (‘black box’). Defendants will be hard pressed to understand that they need to serve a multi-year prison sentence while doubting that the court has sufficiently considered the circumstances of their individual case – which are frequently on an interpersonal level.

A glimpse into the future

These developments are more tangible now than they were just a few years ago. German law in particular, with its generally clear structures and concepts, offers an ideal basis for the digitalised application of the law. The structuring of cases in accordance with the codified law that is at the core of legal studies in Germany is not significantly different from a ‘mind map’ that is part of digital processes. In addition, German law is the foundation of a large number of other legal systems. In conjunction with the now very precise translation programmes. It could therefore become another German exporting success.

About the author

Dr Thomas Wambach is a lawyer based in Hamburg, focusing on commercial litigation. He has vast experience in handling complicated cases in all areas of commercial law. In the last couple of years he specialised in the defence against claims of investors in capital market cases, sample cases and mass litigation. He is a partner and co-chair of the dispute resolution practice at Heuking Kühn Lüer Wojtek. He is regularly recommended by leading handbooks.


Heuking Kühn Lüer Wojtek
Tel: +49 40 35 52 80 29
F: +49 40 35 52 80 80

Return to the Disputes Yearbook 2022 menu