Is the artificial lawyer coming?

Artificial intelligence (AI) is dominating the debates in politics, business and the media. Three offerings stand out: Google's Gemini, Microsoft's Copilot and the current industry leader Chat-GPT from Open AI. You can ask them questions and they respond immediately. They help you write texts, create summaries or even draw pictures.

The quality of the interaction is strongly influenced by the prompts. These are the questions and instructions that users give these chatbots. For example, Chat-GPT can be instructed to take the perspective of a lawyer and create a legal document or a contract for the purchase of a company. Chat-GPT then creates various contractual clauses, which are supplemented or modified with input. A survey by Lexis Nexis shows how frequently AI is already being used, with 36 per cent of lawyers surveyed stating that they use AI in their day-to-day work.

Not really intelligent yet

A recent study by Goldman Sachs even assumes that 44 per cent of the work of lawyers and legal professionals in the USA and Europe will be performed by AI in the future. However, these chatbots are not yet truly intelligent. Even if chatbots spit out words and entire, comprehensible texts and sentences, they are based on a number-based neural network with 176 billion links and options (and the number is growing).

Not everything the AI presents as an answer is always correct, however. Google's Gemini recently attracted attention with depictions of black Wehrmacht soldiers. The phenomenon known as "hallucination" describes incidents in which the AI presents a supposedly convincing result that is not supported by facts.

However, it is not necessary to point out incorrect answers to the chatbot. So far, it has only been learning in training mode, which is only accessible to developers and programmers. Law firms and courts are therefore among the sceptics of advancing digitalisation, as a legal tech survey also shows. Some law firms have their AI solutions customised by specialised programmers. Their own legal pleadings and documentation serve as a database. This is to prevent the chatbot from creating fantasy solutions from the vastness of the internet and the law firms' valuable expertise from ending up on external servers.

Yet, the technology has enormous advantages, regardless of which model is chosen. The time-consuming task of summarising judgements and literature, searching for references and drafting simple contracts can be left to the AI. This leaves more time for the more demanding tasks, which at best results in efficiency gains. In any case, the development of AI will noticeably change the way lawyers and attorneys work.

Success with external content

In addition to the many benefits offered by AI, however, the associated risks and challenges must also be taken into account. With its increasing influence on the world of work, further questions and problems are arising, particularly in copyright, data protection and criminal law. The New York Times is currently suing Open AI and its investor Microsoft. The latter has secured the Open AI algorithm for use in Microsoft products for USD 10 billion. The publisher is claiming various copyright infringements.

The accusation is reminiscent of the years of legal disputes between Google and publishers worldwide. This is because Chat-GPT accesses the knowledge of the publisher of the New York Times, uses its databases with millions of articles and presents them to its own users without paying licence fees. This raises the fundamental question of how the commercialisation of AI should be handled. Do we want to use our input to feed an AI that becomes better and makes its knowledge available to other users? Will the increased use of AI impair the professional skills of lawyers or will it improve the legal working environment in the long term by combining it with technological efficiency? One thing is certain: the development can no longer be stopped - so let's think positively.