Artificial intelligence can now help detect false statements

Researchers have recently developed a tool to help artificial intelligence detect lies by users. Of course, this tool is currently limited to financial areas and it is expected that insurance institutions and banks will start using it soon.

What is the use of lie detection with artificial intelligence?

Artificial intelligence-based software is used in various fields and businesses. For example, such programs can evaluate a person's financial ability to receive a loan or determine the cost of a person's insurance. Algorithms used for this purpose usually use various statistics to do their work, and for this reason, it is possible to manipulate the results by providing wrong information from the users. For example, people can get more loans than they can afford or lower their insurance premiums.

In order to solve this problem, North Carolina University researchers went to edit the algorithms used in existing programs. They determined new training parameters that the AI ​​uses to train itself to evaluate and predict. The mentioned parameters are designed in such a way that they improve the power of artificial intelligence in detecting false data, especially related to people's income.

In the experiments, artificial intelligence trained with new parameters was able to show an acceptable performance in detecting unrealistic data recorded by users. Of course, it goes without saying that still false data with insignificant differences compared to real data cannot be detected by such software and only relatively large differences can be detected.

The researchers of this project have announced that they will soon publish the parameters used in their research publicly so that other developers and experts can also benefit from it. The results of this research have been published in the Journal of Business & Economic Statistics.