Minimizing Technology Bias Through Artificial Intelligence
JAKARTA - Artificial intelligence (AI) is something foreign to some people in Indonesia. The reason is that the term is rarely echoed in daily activities.
Several renewable technologies such as cell phones to vehicle technology are currently involved with artificial intelligence. One of them is on a cellphone that currently has a lot of AI embedded in its sophistication.
Professor of the School of Electrical and Informatics Engineering at the Bandung Institute of Technology (ITB) Bambang Riyanto Trilaksono mentioned three things that cause a social bias towards artificial intelligence.
"What causes bias in artificial intelligence? In my opinion, there are three. The first is training data to build artificial intelligence models", said Bambang in his presentation at the discussion of the National Technology, Information and Communication Council virtually, Thursday, March 25th.
"The second is the artificial intelligence algorithm itself, the third is humans who develop artificial intelligence technology. Each of them can contribute to the bias of artificial intelligence products", he continued.
According to Bambang, overcoming biases on artificial intelligence is not something simple. It is necessary to do a deeper investigation regarding the factors that cause bias, whether in training data, algorithms, or artificial intelligence developers.
He said understanding the data set, including the statistical distribution of whether the data is balanced or not, is also important to do. It is also important to understand the algorithms used to build artificial intelligence models and understand the specific goals of application development.
"As well as understanding social practices in the field where artificial intelligence is created. Then what is also recommended is to obtain input from human experts who should be external from the team that develops artificial intelligence technology", said Bambang.
Biases to artificial intelligence have occurred many times. One of them is when Amazon uses artificial intelligence to recruit employees. In the recruitment process, a bias arises, in which artificial intelligence tends to recruit men rather than women.