Eliott (not his real name) suffered from a benign but massive brain tumor in his frontal cortex. Surgery to remove the tumor was successful, but Eliott’s personality changed dramatically after the surgery.
Before surgery, he was an ordinary, middle-class man with a wife and family, a good job, and no apparent psychological deficits. After the surgery, he could remember his family and his history, could read and understand words, do math, and generally recognize and describe what he saw.
But he began to make very bad choices. He began to trust unsavory people. He made a series of bad investments, losing all of his money. He was unable to connect with his co-workers and his family. He ended up losing his job and his family.
No physician or psychologist Eliott consulted could find anything wrong. He passed all standard cognitive tests and appeared to be psychologically intact. Eventually psychologist Antonio Damasio, a professor at USC, was consulted. After devising a set of new tests, Damasio concluded that the surgery had damaged the portion of Eliott’s brain that controlled emotional responses.
While Eliott’s logical brain functions were unimpaired, Eliott had literally lost the ability to feel emotions. As a result, many life decisions which could not be solved on pure logic became impossible for Eliott to solve at all. He had become an unfeeling logic machine.
Eliott’s story is relevant today as we contemplate our future with so-called “artificial intelligence.” Much like Eliott, AI is an unfeeling logic machine. AI can follow rules and make deductions based on those rules many orders of magnitude faster than humans, based on massive amounts of data from past experiences.
But it cannot feel anything at all. It lacks the self-awareness and the goals and desires that are the bedrock of emotional intelligence. It cannot appreciate beauty or feel love. It has no need for those things.
Just as Eliott lacked the ability to make smart decisions about his future, AI is not equipped to make such decisions either. Which is why AI will never replace humans. It is a very powerful tool for summarizing past events and making connections about the past, information that humans can use to assist us in planning our futures.
But only humans have the creativity and the emotional capacity to understand the implications of what AI discovers.
We need to ban the term “artificial intelligence” from our lexicon. These tools are not “intelligent” in the ways we typically use that term. They are information processors or aggregators, able to summarize vast quantities of information in useful ways. But it requires human intelligence to make good decisions based on that information. True human intelligence requires emotional evaluations that these tools can never do.
AI is a tool. Like any tool, it can be used for good or for evil. Even though a knife can be used as a murder weapon, we don’t ban knives, because knives have social utility in many other contexts. But also like knives, we can impose consequences on those who use the tool to harm others.
We need to understand the ways in which AI might be used to cause harm in society, and craft laws that prevent or punish those consequences.
The AI genie can never be put back into its bottle. But if we are attentive to the harm it can inflict, maybe we can ensure that it is only used to advance society in a positive way.
Rep. Ken Chestek, D-Laramie, is a retired professor of law and the current representative from House District 13 in the Wyoming House of Representatives.