With the rapid development of technology, artificial intelligence tools are becoming central to professional business life. Microsoft is announcing the “Legal Agent” artificial intelligence tool developed specifically for lawyers and legal professionals for the popular word processor Word.

Although this new assistant, which can review contracts, offer regulatory suggestions and detect risky substances, seems quite practical on paper, past artificial intelligence disasters leave some question marks in mind.

What Can Microsoft Legal Agent Do?

According to the information shared by Microsoft, Legal Agent is currently offered to Windows desktop Word users within the scope of the Frontier program in the USA. No separate application or installation is required for this tool, which works integrated with the Copilot infrastructure. The system is mainly designed to review contracts and official documents.

It analyzes a contract clause by clause, highlights risks and liabilities, compares document versions, and monitors changes and offers editing recommendations. While doing all these operations, it completely preserves the original formatting, tables and lists of the document.

On the Google Workspace side, Gemini-supported tools and industry-specific artificial intelligence solutions such as Harvey AI increase competition in the market by offering similar features. However, Microsoft’s integration of this power directly into Word, the platform currently most used by lawyers, gives the company a significant accessibility advantage.

Risk of Artificial Intelligence Hallucinations in the Legal World

Although these features greatly speed up the workflow, the potential of generative artificial intelligence to “hallucinate”, that is, fabricate completely fake data or events, poses great dangers for the legal world. Some lawyers who previously relied entirely on similar tools have faced serious legal problems.

For example, in 2023, two lawyers in New York were fined $5,000 for submitting to the court fake precedent decisions produced by ChatGPT that did not actually exist. Similarly, Donald Trump’s former lawyer Michael Cohen admitted that he used fake court citations created by Google Bard (now known as Gemini) without realizing it. In fact, a French data scientist and lawyer recently announced that he had identified non-existent references produced by artificial intelligence in hundreds of court documents over the past year.

Microsoft Takes Precautions, But the Responsibility Is Human

Microsoft adds various security barriers to the system to prevent both users and the company from falling into a legal nightmare. The legal assistant’s suggestions and quotes are linked directly to the source texts. Thus, the person reviewing the document can instantly verify the source of the presented information before adding it to the official document.

Additionally, there are very clear warnings in the system interface. It is clearly stated that Legal Agent does not provide legal advice, that it may produce inaccurate content, and that the resulting output must be reviewed by a competent legal professional before it is considered reliable. In summary, although artificial intelligence reduces the document review task that used to take hours to minutes, the responsibility for verification and fact checking at the end of the work remains entirely on the shoulders of lawyers.

What do you think about this new artificial intelligence assistant developed by Microsoft for Word and the artificial intelligence mistakes experienced in the past? Don’t forget to express your thoughts in the comments section below.