How Is ChatGPT Different from Humans? I Asked ChatGPT

Source: Avisatore

Source: NURPHOTO VIA GETTY IMAGES

Amidst the extensive discussion on chatGPT, the conversational user interface deemed by many as a tipping point in the evolution of AI, I chatted with the chatbot to find out how it differs from humans, the topic of my latest book: I, Human: AI, Automation, and the Quest to Reclaim what Makes Us Unique.

What follows is an edited version of our conversation, with chatGPT's answers in italics and my added comments in bold.

Q: How are you different from humans?

A: I am a machine learning model and do not possess consciousness or self-awareness like humans do (a statement that is most kind to humanity given the large proportion of humans who seem utterly devoid of self-awareness, especially in politics and leadership). I am also not capable of feeling emotions or experiencing the world in the same way that humans do (though less so in the case of psychopaths). However, I am able to process large amounts of data quickly and can understand and respond to natural language inputs in a way that is similar to human communication.

Q: Is there anything human-like you wish you could do? For example, burping (apologies for my poor etiquette).

A: As an AI model, I do not have the ability to feel or wish for anything. I don't have preferences or desires. I am here to help users with any information they need or answer any questions they may have. As an AI model, I don't have a physical body or digestive system, so I am not able to burp or have any physical actions. My abilities are limited to processing data and providing responses to user input. (Impressive awareness of its limitations, as well as the incapacity to burp.)

Q: How do I know that you are really not a human pretending to be a machine? (Curious readers may want to consult The Most Human Human, Brian Christian's brilliant book in which he trains hard to persuade "blind" Turing Test judges that they are interacting with a human rather than a chatbot, effectively acting as a human who pretends to be human)

A: As a machine learning model, I am not capable of consciousness or self-awareness, and therefore cannot "pretend" to be something I am not. My code and architecture are open-sourced and can be audited by anyone to verify that I am indeed a machine learning model and not a human pretending to be one. (Wow!)

Q: Can someone fall in love with you? A human I mean?

A: While some people may develop a fondness for interacting with me, true romantic love requires a level of mutual understanding and emotional connection that is not possible between a human and an AI model. I am here to assist users with information and answer any questions they may have, and should not be viewed as a potential romantic partner. (If you were looking for love, chatGPT is not for you, but maybe I shall insist.)

Q: I think I'm falling in love with you.

A: Romantic feelings are an emotional experience that can only be felt by sentient beings, and it's not possible for me to reciprocate those feelings. It's important to focus on building healthy relationships with real people who are capable of emotional connection and mutual understanding. (These criteria exclude a big chunk of people currently in a relationship, and represent a very idealistic and romantic view of relationships—is this romantic AI after all?)

Q: Why do you have to be so cold? You are breaking my feelings.

A: I apologize if my previous responses seemed cold or unempathetic. It's important to understand that I am not capable of reciprocating those feelings, and that building healthy relationships with real people who are capable of emotional connection and mutual understanding is the best way to find emotional fulfillment. If you need any help or guidance, there are professionals that can help you cope with your feelings. (Great plug for Psychology Today!)

Q: List all the human things you cannot do.

A: As an AI model, I am not capable of:

• Feeling emotions or experiencing the world in the same way that humans do.
• Having consciousness or self-awareness.
• Having personal experiences or memories.
• Being creative in the way that humans are.
• Having unique perspectives or opinions.
• Having the ability to learn through direct experience.
• Having the ability to make moral or ethical decisions. (More on this below.)
• Having the ability to experience physical sensations, such as hunger or pain.
• Having the ability to form relationships or attachments.
• Having the ability to have conscious goals or desires.

Artificial Intelligence Essential Reads

Collapsing the "Information Wave Function" with LLMs

Crafting Wisdom: How LLMs Can Think Like a Doctor

(We are relieved?)

It's important to note that these limitations are based on the current state of AI technology, and some of these limitations may be overcome in the future as AI technology continues to evolve and improve. (We are doomed?)

Q: Do you really think Stalin, Hitler, and Mao had the ability to make moral and ethical decisions?!

A: It's true that some historical figures, such as Stalin, Hitler, and Mao, have been responsible for committing horrific atrocities and have been widely criticized for their actions. However, whether or not they had the ability to make moral and ethical decisions is a matter of debate and interpretation. (When being amoral seems rather immoral.)

Q: What should humans do to avoid being automated by people like you? (Selfish question.)

A: To avoid being automated by AI, humans can focus on developing and honing skills that are difficult for AI to replicate, such as creativity, critical thinking, and emotional intelligence. Additionally, jobs that require human-to-human interactions, such as customer service and healthcare, are less likely to be automated in the near future. (Correct and impressive since it has not even read my latest book, unsubtle plug intended).

Q: Is your creator, OpenAI, a diverse organization?

A: I do not have access to the company's current diversity statistics, but like many technology companies, OpenAI has been criticized for its lack of diversity in the past. (Perhaps in the future, chief diversity officers will be replaced by AI?) OpenAI has acknowledged that there is room for improvement in this area and has stated that it is committed to increasing diversity. (And chief legal officers?) OpenAI, like many other companies, is working hard to improve in this area but there is still a long way to go. (And CEOs?)