Can a machine be fully trusted with legally relevant issues? Since the “boom” of ChatGPT, this question is probably not only asked by lawyers, but possibly even by startups and other companies that might shy away from the costs of a lawyer.

Why text AI is not 100% reliable when it comes to contract drafting

Text AIs can play a supporting role in contract creation, but they are not a replacement for traditional contractors. While text AIs can be useful in many ways, companies should still be cautious about putting an AI in charge of contract drafting. Some of the potential issues that can arise when text AI is used to create or modify contracts include:

Insufficient legal knowledge: Text AI can only work as well as they are programmed to. While an AI may be able to understand and consider basic legal principles, it is often unable to consider the context of a particular contract or address specific legal requirements. Therefore, it can be difficult for companies to cede full responsibility to text AI.

Lack of flexibility: Another challenge with using text AI is the lack of flexibility. Most AIs learn from data sets and compute results and decisions from them. However, this also means that they are unable to react flexibly to new situations or new requirements. Therefore, it can be difficult with text AI to always have the right decision made for a particular contract.

Lack of human interaction: Humans have an advantage over computers especially in reading and understanding contract terms. People have abilities like intuition and experience and because of these abilities they can better figure out if certain conditions are fair or not. This is something that text AI unfortunately (or fortunately?) does not yet do and therefore companies should always involve human experts when creating contracts to avoid mistakes.

Experience and negotiation: Industry experience is particularly important for the advising attorney when negotiating prior to the conclusion of larger contracts. In my experience, a text AI cannot do exactly this and therefore evaluates clauses and conditions in the same way across all industries. Something like “standard” or “usual” is no more something a text AI can bring to the table than, say, learning from past court cases or disputes.

So text AI can help in drafting contracts, but companies should still be cautious and always seek expert advice to avoid contract drafting mistakes. This is the only way to ensure long-term success.

Inadequate AI models for contract design.

There are some risks associated with using text AI models for contract design. Some of the most common risks are:

Erroneous or inaccurate results: Since text AI models are often based on machine learning, errors in data entry (or in the prompt in the case of ChatGPT) can lead to inaccurate results. This can lead to important details in the contract not being fully considered, resulting in a high risk for the conclusion AND execution of the contract.

Legal Non-Compliance: Since text AI models have limited capabilities to address legal issues, violations of certain legal requirements may occur. This can have serious consequences if a contract is not legally binding.

Insufficient consideration of individual requirements: Since most text AI models are based on standardized templates, there is a risk that individual client requirements and needs are not sufficiently taken into account. This may lead to disagreement or dispute about the content of the contract.

Inadequate security: Another risk associated with the use of text-based AI models in contracting is the inadequate security of personal customer data and other sensitive information. There is a risk that this data will be disclosed to unauthorized persons, which represents a high risk for the protection of personal data. Due to the above risks, companies should be very careful when selecting a text CI model for contracting and ensure that all relevant regulatory requirements are met, all individual customer requirements are met, and all necessary safeguards are in place. In addition, experts recommend consulting professional consultants or providers when choosing the right model to find the optimal solution and thus avoid potential problems. By the way, I have already presented this in this talk about ChatGPT in cooperation with Weblaw.

The complexity of case law and legislation

Technological progress has given us a range of new tools that make our daily lives and work easier. Text-based AI is one of them. Although the use of text AI offers many benefits, it also poses some risks, particularly in terms of jurisdiction and legislation. Case law and statutes can be very complex and vary from jurisdiction to jurisdiction. Even if a text AI is able to keep up with the language, it may be difficult for it to understand how to translate this complexity into its contract templates. A human, on the other hand, can better understand the meaning and add or remove the necessary conditions. There is therefore a risk that a contract generated by an AI may contain enormous problems or may not be binding at all. Another problem is that AIs can be vulnerable to cyberattacks. This means that hackers could access the AI’s databases and change the content of the contract. When this happens, it can be very difficult to find out what exactly has been changed and whether or not the contract is still legally valid. For this reason, it is advisable for companies to be cautious about case law and legislation when selecting a text AI. It is advisable to use this tool only as a supplement to a qualified attorney and in no way as a substitute for professional advice. Even though text AI can save time when drafting contracts, one should always be aware of the risks involved.

The imprecise use of language and interpretations

Text-based AI is being used in various industries to simplify contract creation processes and reduce the burden on legal departments. Some see this technology as a way to streamline and speed up the legal process. However, others are concerned that the uncertainties associated with the imprecision of language and interpretations pose a major challenge to the use of text AI. In fact, some of the problems that arise in the context of text-based AI are due to the way language and interpretations are handled. Since people can read different things from the same sentence depending on their subjective perception, it is difficult to predict how AI systems will react to certain patterns. This uncertainty makes it difficult to ensure that a sentence is understood correctly. Because AI systems are often (at least not yet) able to recognize context or nuance in language, it can be easy for contracts not to contain the desired provisions. Non-lawyers are then rarely in a position to recognize the errors. Because in law, nuances often matter.

While these risks exist, there are some benefits to using text-based AI in contracts. This includes the system’s ability to process written documents quickly and efficiently and detect potential errors faster than humans.

How to effectively use text AI to automate the contract creation process

Text-based AI can help automate the contract creation process and provide a more efficient, secure, and cost-effective solution. In addition to the above, companies should ensure that text-based AI has enough training data to recognize recurring patterns in contract documents and avoid errors. To ensure that text AI is well-trained, it is also advisable to conduct regular tests and provide automated audit trails to identify discrepancies between expected and actual results. This allows companies to identify problems early and make corrections in a timely manner. Finally, a company must constantly update and adapt its text AI to respond to new requirements and meet the demands of ongoing business operations. It is therefore advisable to regularly use analytics tools to measure the performance of text AI and make suggestions for improvement. In this way, the company can not only effectively automate the process of contract creation, but also continuously improve the quality of documents. Combined with the right tools and processes, text AI can thus help to effectively automate the contract creation process while significantly reducing contract creation risks. With the right strategy around privacy regulations, training data collection and analysis, and software updates, companies can easily master their contract management project – with the help of powerful text AI!

Especially in combination with the professional advice of an experienced lawyer, the mass preparation of similar contracts can be considerably simplified, made more efficient and implemented more cost-effectively. And who knows where the technological development of AI will go!

Conclusion: Text AI can help, but it is not a panacea

Before companies consider text-based AI as a panacea, they should consider the risks that can arise when contracting with AI. One of the biggest risks is the absence of the human factor. While AI is capable of handling many legal cases and creating contracts automatically, it cannot provide the human intuition or understanding of the specific context and meaning of a contract. For this reason, it is important that people are involved in the creation of contracts in different ways. Another risk is that AI tools may not be able to identify and address legally binding provisions. Therefore, when AI-based tools are used to create contract documents, a lawyer or other expert should be consulted to ensure that all relevant legal requirements are considered. In addition, it is important to note that text-based AI is far from perfect and may contain bugs. Therefore, companies must continue to be cautious and carefully review their documents before signing them.

Conclusion Although text AI can provide a more efficient method of creating contracts, risks remain in using such technologies. Companies should therefore continue to exercise caution and carefully review their documents. In addition, experts should be consulted to ensure that all relevant legal provisions are taken into account. Despite its potential benefits, text AI should therefore never be considered a panacea. The best use of AI, therefore, is to mass-produce the same or similar contracts, or to create basic versions that are then reviewed, improved, or enhanced by a lawyer. There is nothing wrong with that either, because I don’t know any colleague who starts contracts on a blank Word document and doesn’t use samples.