The air crackled with anticipation at the German Institution of Arbitration (DIS) Spring Event 2025 on May 6th and 7th, as attendees delved into a transformative question: Could Artificial Intelligence truly revolutionize the world of dispute resolution? For those eager to understand the seismic shifts on the horizon, here are seven key takeaways from this insightful gathering:
Takeaway #1: Enormous interest and positive sentiment regarding AI
The DIS Spring Event revealed significant interest and positive sentiment regarding the integration of AI in arbitration. This was reflected in the large number of participants and the high level of attendance during the various program items. Further, a survey was conducted using Mentimeter in order to record the opinions of the participants on the use of AI in dispute resolution. The results of this survey showed a clearly positive trend in the assessment of AI technology. A majority of the participants expressed the expectation that the use of AI would improve the quality of dispute resolution. This optimistic attitude indicates a general openness and a positive basic feeling within the arbitration community towards the integration of AI.
Takeaway #2: Amara’s Law: The time is now
During the event, Jörg Risse (Baker McKenzie) presented Amara’s Law, which states that the impact of a new technology is overestimated in the short term, but vastly underestimated in the long term. He illustrated this using a typical technology curve, which leads from initial hype through a valley of tears and a path of enlightenment to exponential development. According to his assessment, we are currently at the point of exponential development in the field of AI. For a long time, legal tech and similar topics were discussed intensively, but without any real breakthroughs being achieved, which led to a certain hopelessness. Now, however, rapid developments can be observed on a weekly or monthly basis, triggering a new wave of confidence.
Takeaway #3: Deep search can do amazing things
Another highlight of the event was the demonstration of Gemini Deep Search in the context of arbitrator searches. Daniel Schnabl (Freshfields) and Nathalia Schomerus (Legora) demonstrated how they entered criteria for the selection of arbitrators into Gemini (an AI chatbot developed by Google). The deep search took a few minutes, but delivered impressive results and detailed evaluations. Particularly noteworthy was the fact that the system also revealed the reasoning behind its suggestions. However, a humorous but instructive example showed that the results must always be critically examined: Among other things, the system suggested a referee from the German Football Association, who would obviously not be ideally suited to a real arbitration procedure. It also became clear that caution is required when prompting. For example, if a user only asks for “Schiedsrichter” in German, Gemini will only provide male arbitrators.
Takeaway #4: We are still skeptical about decision-making by AI
There was clear skepticism at the event with regard to decision-making by AI. There was intense discussion about whether it will be possible in the future for AI to make decisions in legal disputes instead of humans.
Jörg Risse presented two fictitious examples of this. The first scenario showed an innovative chairperson who had two AI systems deliver a preliminary assessment of the case. The second example was a robo-judge procedure for air passenger rights disputes (Fluggastrechte). In such proceedings, an AI could decide on the lump sums claimed. If a party does not agree with the decision, it can appeal against the AI’s decision and then initiate “normal” court proceedings. The promised benefit: a reduction in the workload of the judiciary – illustrated by a reduction from 48 judges at the local court in Frankfurt am Main, who currently deal with passenger rights disputes, to just 6.
Various risks were raised during the discussion on robo-judges. One important point was the assessment of evidence: How can AI make an assessment of evidence if the facts are not clear? Constitutional concerns were also raised, in particular the question of whether decisions should be made exclusively by humans. Reference was also made to the EU AI Act, which classifies AI as a high-risk system in decision-making. Finally, the question was raised as to whether decisions made by AI would even be enforceable under the New York Convention.
The reluctance towards AI decision-making reflects fundamental concerns about the rule of law, fairness and the role of human judgment in legal proceedings. Legal decision-making often requires nuanced interpretation, consideration of context and the application of legal principles, which still poses a significant challenge for current AI systems.
Takeaway #5: Significantly less skepticism about AI-supported decision-making
In contrast, there was much less skepticism about the proposal to use AI as a digital assistant. This proposal was put forward by Florian Diekmann (judge and president of the Regional Court Rottweil). The idea envisages AI carrying out an initial examination of the case and issuing a vote. This vote would then be presented to the judge together with a statement of reasons in order to support him or her in the decision-making process.
A similar approach was also discussed for arbitration proceedings, where AI could act as a kind of sparring partner or assistant to the tribunal, similar to the secretary of the tribunal. The issue of confidentiality was not seen as an insurmountable risk in this context. It was argued that confidentiality is not mandatory in arbitration proceedings either and can be waived by the parties. Furthermore, even without the use of AI, there is always a risk that confidentiality will be breached. A careful selection of AI systems could help to protect confidentiality in the best possible way.
The greater acceptance of AI as a digital reporter compared to decision making suggests that there is a consensus to use AI to support human skills rather than replace them entirely in the legal process.
Takeaway #6: There is a real client need for AI usage
Franziska Fuchs (Robert Bosch) reported that Robert Bosch is working intensively with AI and is already realizing enormous savings potential. For example, AI-supported summaries of depositions have achieved savings of up to 90% compared to manual processing. Quality has also been improved, as human processors tend to make more mistakes. This also raises questions with regard to the Business Judgment Rule: When is it still appropriate to bring in external consultants when AI could potentially perform the same task more cost-effectively and efficiently?
According to Ms. Fuchs, companies such as Bosch will in the future provide their external consultants with AI-based preliminary products and assessments on which they could then build. This suggestion led to skepticism in the audience regarding the question of liability. However, Bosch’s message was clear: law firms must work together and openly with their clients on AI solutions in order to exploit the potential for savings, efficiency gains and quality improvements. It is to be expected that many specific AI-related requirements will come from clients in the future.
Takeaway #7: How we do our work will change
Participants generally agreed that AI adoption will lead to changes in legal work practices. Experienced professionals will need to adapt to new technologies, potentially altering established routines. Younger associates will increasingly delegate tasks to AI systems, requiring the development of skills in verifying AI outputs. Psychological aspects, such as automation bias and memory blindness, as highlighted by Ragnar Harbst in Global Arbitration News, were noted. Over-reliance on AI summaries could lead to the overlooking of inaccuracies and potential memory distortion. Lawyers will need to critically evaluate AI outputs against original documents and remain aware of AI limitations to mitigate these risks.
Outlook
In summary, the DIS Spring Event underscored the considerable interest in and generally positive assessment of AI’s role in arbitration. While legitimate concerns, particularly regarding data protection, exist, the increasing relevance and advantages of AI are widely acknowledged. The central message from the event is the need for legal professionals to adapt to these evolving technologies, collaborate with clients on AI solutions, and approach the associated challenges with a proactive and open perspective. The prevailing mindset should be one of seeking solutions and embracing new possibilities (“Yes, and…” instead of “Yes, but…”).
