Perhaps I am mistaken. But it is my impression that there’s a normative aspect of AI that needs to be addressed and hasn’t been. This aspect is a certain normative schema of creating AI minds.
When most people think of AI, such as myself, they think of incredibly intelligent—in the cognitive sense—beings; beings that may or may not be smarter than the average human. These beings often lack certain qualities, such as wisdom, emotional intelligence and the ability to love. In short, we think of the stereotypical human “genius,” who may bumble about, not tying their shoes but producing great theories of the universe.
I want to question the normative aspect of creating a certain type of mind in AI.
As we know, there’s diversity in the human species. Humans can have autism, Down’s Syndrome, schizophrenia. The fact that these so-called disorders are not inspected more by AI developers is telling. And it purports to create a new society—an AI society—without such “disabilities.”
It has been argued by many that the differences in minds—often referred to as “neurodiversity”—is just as beautiful as other forms of human diversity. And these differences give us strength within our communities.
No doubt, it would be difficult to create an AI with a condition like autism and there would be criticism all around such a project—from all sides. But boldly going into that territory nevertheless is perhaps better than ignoring such human differences when it comes to developing AI.
I want to challenge AI developers to look beyond their own current understanding of what they are producing and aim to create artificial intelligence with flaws, deficiencies, vulnerabilities and, yes, disorders. This is perhaps much more challenging than creating the standard AI, it’s true. But erasing the differences in humans while creating AI is perhaps worse.
We cannot deny that most bots created fit into a certain category. This inflexible understanding of the human mind only wipes away the beauty that is humankind. As you think about the characteristics you program and develop into your AI, think about the diversity that lay behind your assumptions.