Artificial Intelligence (AIs)| Fair or not?

Nowadays, the most buzzing word is AI (Artificial Intelligence). That means a machine can predict or can decide like a human with its intelligence; Artificial intelligence. This technology spread over all sectors of our life. As it helps to make the critical decision with the shortest path, it helps to define diseases, predicts medicine against the conditions, control traffic system, etc. All these are for people's good. So there is no way to avoid AI machines. But crisis occurs when the robots decide by its intelligence, and it goes against our life or our safety. Then what will happen? Are we ready to face the problems, or can we solve them? Don't worry, scientists are aware of it. They also think about it.

The University of Massachusetts Amherst researchers is working to solve the problem and find a way to prevent the worry. They build a framework that can answer all the questions above. They named this "Undesirable behavior" in intelligence machine. Can it solve the problem, and is the research is successful?

In general, an Algorithm controls the machine which works with AI. And it's tough to control the Algorithm's behavior. On the other hand, it's essential to ensure the fairness of AI machines for us, or it will harm our lives, which will be immeasurable.

The researcher of the University of Massachusetts Amherst designs an Algorithm for their framework. They named the Algorithm as the "Sheldonian Algorithm." They make a bid that this Algorithm will ensure a controlled behavior of AI machines, which is excellent and safe for us. So what is the change?

The trending AI is very simple. As the designers design an Algorithm and user provides data, the Algorithm makes decision/prediction according to the provided data only. In this way, the machine can make a decision that may not be right. But the new framework and Algorithm prevent the problem. This framework contains a design with a two-part interface and Algorithm. In the interface part, the user can define the undesirable behavior. The Algorithm decides according to the provided data and avoiding unwanted behavior. Such they told about using the Sheldonian Algorithm for diabetes treatment. Here the undesirable behavior is low blood sugar or hypoglycemia. Now the machine, when controls the insulin, will pump it will avoid hypoglycemia increase frequently. This feature is a unique one, and the earlier AI machines did not have this. This new machines will decide fairly for people's welfare. 

The research is quite a success, but there are many ways to invent new features and make the Al machine more recline to people. The researcher of The University of Massachusetts Amherst Tomas Philip says:

"We believe there's massive room for improvement in this area. Even with our algorithms made of simple components, we obtained impressive results. We hope that machine learning researchers will go on to develop new and more sophisticated algorithms using our framework, which can be used responsibly for applications where machine learning used to be considered too risky.