ANKARA (AA): Artificial intelligence (AI) having a say in issues that are vital for humans may no longer be merely science fiction. Even in as a critical field as law, it has started to be used in pilot programs in some places around the world.
There is debate whether it is ethical for algorithms that mimic human behavior to have a voice even in courtroom decisions.
In the field of law, artificial intelligence, which is believed to help speed up litigation and automate routine work, is fielding various pilot applications in different parts of the world, for example in China, Estonia and Malaysia.
With “robot judges” evaluating small cases in Estonia, robot mediators in Canada, artificial intelligence judges in China, and an artificial intelligence judicial system in Malaysia, it is now possible to see algorithms in the justice system.
There are certain principles regarding the moral control of AI, Professor Ahmet Ulvi Turkbag, a lecturer at Istanbul Medipol University’s Law School, told Anadolu.
“The most important of these is that AI should be transparent. It must be absolutely controllable. Because if we don’t know why a decision is made, we cannot make a judgment about the correctness of that decision. This can lead to very dangerous consequences,” said Turkbag.
Saying that AI has the power to make surprising decisions and therefore the decisions made by algorithms should be accessible to humans, Turkbag argued that this can be achieved with small programs called “subroutines.”
He said important court rulings made by algorithms should also be auditable by human intelligence.
Manipulation, privacy concerns
Some experts worry that the algorithms are “deceptive and pose a risk to privacy and public safety.”
The non-profit Center for Artificial Intelligence and Digital Policy (CAIDP) in the US has applied for the Federal Trade Commission to stop the use of GPT-4, the new version of the OpenAI company’s artificial intelligence robot ChatGPT.
Some industry experts are concerned about human manipulation of computer technology.
“AI should not be manipulated, this is very important. You asked the AI to save a human, and this person also has a pet, and AI should not kill an animal while saving a man,” said Turkbag.
AI decisions face a higher authority
Turkbag said that, hypothetically, if a decision made by AI is appealed and brought to a higher court, the case should be taken over by human intelligence.
“Even if we accept artificial intelligence in the first stage, it should definitely go to humans if it faces objections, the logic of law requires it,” Turkbag said, adding that AI should conduct a large-scale database scan depending on the importance of the case.