If data is the new oil, Artificial Intelligence is the combustion engine of our era. In 2007, five of the world’s ten most valuable companies were linked to the oil industry. Ten years later, in 2017, seven of the top ten positions were occupied by technology companies. Would anyone care to guess what the ratio will be in 2027? Experts have described this shift as the era of “data as the new oil,” to the extent that Artificial Intelligence (AI) has been compared to the transformative impact of the machines that fueled the Industrial Revolution.
This revolution is also arriving like a steamroller in the legal profession. For the first time, the Justice in Numbers 2026 report, released by Brazil’s National Council of Justice (CNJ), revealed that 428 million legal cases are stored on the Sinapses platform, a tool developed to train, share, distribute, and audit AI models used by Brazilian courts. These millions of cases allow machines to learn how to automate workflows, classify documents, and identify patterns.
When we examine institutions connected to the judiciary, we find that 94% of them already use some form of Artificial Intelligence. Moreover, just one month ago, Brazil’s Superior Court of Justice (STJ) amended its internal rules to require structured summaries in legal filings, making them easier for automated systems to process and analyze.
Globally, ambitious plans are underway. The United Arab Emirates has committed to becoming the world’s first “AI-native government” by 2027, integrating AI across the entire public sector.
AI in the Justice System
In May of this year, Anthropic, the company behind Claude, launched legal-focused functionalities directly on its platform, including contract review, compliance verification, and legal risk assessment tools. However, these features were designed primarily for the U.S. legal system, not for Brazilian law.
Some professionals may be using these tools without fully understanding this limitation, and that is precisely where the challenge lies: technology is advancing faster than users’ ability to understand where it works effectively and where it falls short.
There are also numerous cases in which court submissions have contained AI-generated content with fabricated information, commonly referred to as “hallucinations” produced by large language models. It is worth emphasizing that the problem is not the use of AI itself, but rather its use without proper judgment, review, and, above all, professional responsibility.
According to the TIC Government 2025 report, only 59% of federal government agencies provide training for AI use. In the private sector, similar questions must be raised regarding professional qualification and AI literacy, and they must be addressed urgently. After all, when something goes wrong, the attorney remains accountable.
Millions of Lawsuits and Millions of Reasons to Discuss AI
The year 2025 recorded the highest level of judicial demand in Brazil’s historical series, with 1.4 million more new cases than in 2024.
The legal news portal Migalhas conducted an interesting forecasting exercise, analyzing the growth in litigation following the launch of ChatGPT at the end of 2022. If the pace observed over the last three years continues, the number of new cases filed before Brazil’s Labor Courts could reach approximately 9.55 million by 2030. Cases before the Small Claims Courts could rise to approximately 17.99 million, while the State Courts could receive around 18.25 million new lawsuits by 2030.
These figures are, of course, projections, and the portal itself acknowledges that important variables are still missing for any definitive conclusion. Even so, the scenario is difficult to ignore. One can only imagine the impact on a judiciary that already faces substantial levels of abusive and predatory litigation, and how generative AI tools may further facilitate the activities of bad-faith litigants.
Nearly a decade ago, in 2017, John Cryan, then CEO of Deutsche Bank, stated: “In our bank, we have people behaving like robots. Tomorrow, we will have robots behaving like people. Whether we, as an institution, choose to participate in these changes or not is irrelevant. They will happen.”
Cryan was right. His only mistake was suggesting that participation was optional.
Collectively, we are experiencing a phase of technological adoption without the maturity necessary to use these tools responsibly. What is lacking is technical accountability regarding a technology we are still learning to master, along with the honesty to admit that this revolution has arrived with overwhelming force.
Artificial Intelligence has already read the law. But can the law read AI?
Available at: https://www.migalhas.com.br/depeso/462227/a-ia-ja-leu-o-direito-mas-o-direito-sabe-ler-a-ia
Autor: Sergio Luiz Bernardelli Junior • email: sergio.bernardelli@ernestoborges.com.br
