What is the One Word That Separates Humans From Artificial Intelligence?

As artificial intelligence gradually enters many areas of life and begins to gain more importance, researchers are trying to find the word that will distinguish between humans and artificial intelligence.

What would happen if we had to distinguish between a human and a robot with artificial intelligence without seeing or hearing, and decide which one is human with a single word?

Which word characterizes a person?

By conducting research on this subject, scientists both clarify our assumptions about artificial intelligence and reveal some surprising information about the functioning of the human brain.

Artificial intelligence is now used in the fields of automatic chat and language production, as well as writing the long texts we encounter in daily life. How can we understand whether the customer representative we correspond with online is a real person or a product of an algorithm? Or how do we decide whether a fictional story is machine- or human-generated? We must be prepared for this, as communicative artificial intelligence is now beyond a theoretical possibility.

John McCoy, from the research team at the Massachusetts Institute of Technology (MIT), questioned the possibility of using a test developed by British scientist Alan Turing in 1950 to distinguish machine intelligence from human intelligence.

Accordingly, a standard chat interface is opened to each referee via the computer. They have to guess without knowing whether they are having a text conversation with a real person or with a computer program equipped with artificial intelligence. If the bot chat can convince a certain number of referees, it passes the Turing test.

McCoy and his team are working on whether a single word can be decisive in this test. "In this case, the question arose as to which word is unique to humans," he says. The results of the research were published in the Journal of Experimental Social Psychology.

In the first experiment, McCoy and his colleague Tomer Ullman asked 1,000 people to suggest a word that could be unique to humans and tried to find what the resulting words had in common.

The words in the top 10 and the number of recommenders were as follows:

•Love (134)

•Compassion (33)

•Human (30)

•To please (25)

•Mercy (18)

•Empathy (17)

•Emotion (14)

•Robot (13)

•Humanity (11)

•Live (9)

“It was interesting how much overlap there was between the words people were suggesting,” says McCoy. "While there were so many words to choose from the dictionary, people chose words that provided such a large overlap."

10 percent of participants chose the word "love" out of hundreds of thousands of possibilities. The total share of the words in the top four was 25 percent.

Among general topics, bodily functions (like "going to the bathroom"), faith and forgiveness (like "compassion" and "hope"), emotions (like "empathy"), and food (like "bananas") were the most popular categories.

In their second experiment, McCoy and Ullman tried to understand how other people evaluated the words revealed in the first experiment.

They formed pairs of different combinations of the most frequently mentioned words (such as "human" and "love") and asked other subjects to guess which one was suggested by the human and which was suggested by the computer.

As in the first experiment, the word "love" was found to be the word most attributed to humans. However, the word that came up the most among the available options was the word "shit." However, as a common characteristic of humanity, an expression of emotion was seen as much more decisive.

Words that people used outside of the dictionary definition were also among the commonly suggested ones.

The reason for this can be seen as a reflection of the current level of artificial intelligence. While computer programs and robots can now write simple descriptive sentences and even short stories, they have difficulty with human humor and sarcasm. In order to choose humor and the words used in this context correctly, it is necessary to understand the subject and know the cultural connections hidden in each word.

Besides these speculations, McCoy believes that this experiment offers an important tool in understanding people's inherent assumptions about other groups. For example, what word would you choose to prove that you are a woman? Or with what single word can you prove that you are French or a socialist? In these cases, words that each group member knows within their own group and that outsiders may misunderstand or disregard are chosen.

McCoy believes this simple question prompts people to think deeply about the distinction between humans and artificial intelligence and the communication between the two.

McCoy says that his favorite among the words that emerged in the experiment was "ummm…" and that he considered it a "clever suggestion."