Can We Trust AI to Be a Fair Judge in Court?

No matter where you live in the world, the courts could probably be doing a better job at dealing with small disputes and other more serious cases. Judges usually have a large backlog of cases that they cannot manage to finish because of the constant inflow of new cases that are higher in priority. 

Estonia seems to have found a way to solve this everlasting problem. They intend to do it with an AI that would serve as a judge and adjudicate small disputes and claims of less than €7,000. But can we trust the AI to be a fair judge in court? What if it makes a mistake and we are wrongfully sentenced? They’ve thought about this as well.

Is an AI Judge a Good Idea?

Estonia has a reputation of being a “digital country” for almost 20 years, after they began digitizing government services in the early 2000s. Today, all of its 1.3 million residents use national ID cards and complete most services online, including tax filing and voting. Therefore, this idea might work well in a country that is used to efficient digital solutions.

This project is still in its early stages and will begin by dealing with contract disputes. Basically, the two sides will upload all documents and share all relevant information, after which the AI will give its opinion and make a decision. 

Of course, this decision is not final. If one of the parties is not content with it, they can file an appeal immediately and a human judge will take over the case. 

As is always the case with artificial intelligence, his AI will also become better as time goes by. In the future, it might even take more serious cases for human judges, to decrease the percentage of mistakes and wrongful sentencing.

Previous AI Projects Show Promising Results

The “Robot Judge” is not the only project including artificial intelligence created with the intent to make the government as lean as possible. 

For example, several years ago, the government started using an AI that would replace inspectors who check on farmers who receive subsidies to cut their hay fields each year. The AI does this by comparing satellite images taken by the European Space Agency. These are fed into an algorithm developed by the Tartu Observatory, in Finland. 

The AI analyzes the pixels in the images to determine whether hay has been cut or not. If the fields are being grazed by cattle or just partially cut, a human inspector will drive to the destination in question to check themselves. 

In case a farmer still has not cut their field, the AI will send them a reminder two weeks before the deadline, along with a satellite image of their uncut field. 

So far, this system has saved around €665,000 in its first years because of the decrease in the number of times the inspectors had to drive to the location to check on farmers.