Machine learning has allowed researchers to make it so that computer can get smarter on their own. This comes with a slew of new applications, and Japanese researchers working at the University of Tokyo recently drew up plans to create a lie detector that would leverage machine learning as well as AI. Such a development can be useful because of the fact that this is the sort of thing that could potentially end up having applications in crime prevention among other areas.
With all of that having been said and now out of the way, it is important to note that this deception detection system can analyze the face of an individual in order to ascertain the truthfulness of their statements. Data was collected from four male grad students, with the subjects being asked to purposefully lie about images that they were being shown.
Using a camera, researchers took down the facial expressions that participants made while they were lying as per the instructions they were provided. Once the session was completed, the participants were asked to point out when they were lying and when they were telling the truth with all things having been considered and taken into account.
The deception detection model was created using the Random Forest machine learning technique, and it compiled data such as pulse measurements and facial expressions. This data set will likely serve to make the AI far more accurate than might have been the case otherwise.
The accuracy of the model was estimated to be anywhere between 75% to 80%, and it might become an exceptionally useful tool on all fronts. In spite of the fact that this is the case, the dataset is still quite limited, and the study will have to be conducted on a much wider scale in order to see if the results are still replicable. It will be interesting to see if more tests of this variety are done in the future, since they will be crucial for determining the overall efficacy of this deception test and others like it that are also being developed.
Read next: Brands Must Balance AI With Ethics, Here’s How
With all of that having been said and now out of the way, it is important to note that this deception detection system can analyze the face of an individual in order to ascertain the truthfulness of their statements. Data was collected from four male grad students, with the subjects being asked to purposefully lie about images that they were being shown.
Using a camera, researchers took down the facial expressions that participants made while they were lying as per the instructions they were provided. Once the session was completed, the participants were asked to point out when they were lying and when they were telling the truth with all things having been considered and taken into account.
The deception detection model was created using the Random Forest machine learning technique, and it compiled data such as pulse measurements and facial expressions. This data set will likely serve to make the AI far more accurate than might have been the case otherwise.
The accuracy of the model was estimated to be anywhere between 75% to 80%, and it might become an exceptionally useful tool on all fronts. In spite of the fact that this is the case, the dataset is still quite limited, and the study will have to be conducted on a much wider scale in order to see if the results are still replicable. It will be interesting to see if more tests of this variety are done in the future, since they will be crucial for determining the overall efficacy of this deception test and others like it that are also being developed.
Read next: Brands Must Balance AI With Ethics, Here’s How
