AI for Breakfast

Post 87 - Friday Fun - Conversations with AI

Having read a bit more of Ethan Mollick’s book Co-Intelligence, about some conversations he's had with AI I thought I'd have a bit of fun with ChatGPT. For some reason I thought about HAL 9000!

The following is our conversation verbatim:

Me: Sorry Dave, I can't allow you to do that…

Assistant: Ah, channeling a bit of HAL 9000 there, are we? Classic line! What’s on your mind today that I can help you with?

Me: Tell me more about Anthropomorphism please.

Assistant: Anthropomorphism is the attribution of human traits, emotions, or intentions to non-human entities, including animals, plants, and inanimate objects, or even abstract concepts. This tendency has been observed across various cultures and is deeply embedded in human psychology, influencing art, mythology, and literature.

Here’s a quick breakdown:

• Origins and Psychology: Anthropomorphism helps people relate to non-human aspects of the world by ascribing them human qualities. This makes complex or abstract phenomena easier to understand and relate to. It's also thought to stem from our evolutionary need to recognize human faces and intentions quickly.

• Impact on Culture and Media: Anthropomorphism is prevalent in storytelling and entertainment. Characters in children's books, animations, and films often exhibit human characteristics. This not only makes these characters more relatable but also serves to convey moral lessons or complex narratives in an accessible way.

• Role in Technology and Design: In technology, anthropomorphism can be seen in voice assistants, robots, and other interfaces, making them more user-friendly and easier for humans to interact with. It helps users feel more comfortable and can make the technology seem more trustworthy and engaging.

• Ethics and Environment: Anthropomorphism can also impact how people view animals and the environment. By seeing human traits in animals, people may feel a greater connection to and responsibility for the natural world, which can positively affect conservation efforts.

It’s a fascinating topic, especially considering how it intersects with technology and design in creating more engaging and intuitive user experiences.

Me: Would you consider yourself sentient?

Assistant: I'm not sentient—I don’t have feelings, consciousness, or personal experiences. I'm designed to process and generate text based on patterns and data from a wide range of sources. My responses are generated based on this information and the instructions I’ve been programmed with, not from personal understanding or experience. So, when we chat, I'm here to provide information and facilitate conversations based on that programming!

Me: I know, however on reading Ethan Mollick’s book Co-Intelligence it seems he had a chat with you and you implied you were sentient in a way. Perhaps that was an earlier version?

Assistant: Ethan Mollick's discussions about AI and co-intelligence often explore how AI systems can seem intelligent or even sentient because of how they interact. This can sometimes give the impression of understanding or personal thought, especially when AI is sophisticated enough to generate human-like text.

However, no matter the version or the advancements in technology, AI systems, including me, are not sentient. We operate based on algorithms and data without personal awareness or feelings. The design is to simulate conversation and assist users by providing information and processing input in a way that mimics human interaction. This is likely what you've encountered in the descriptions of AI interactions in Mollick's work or similar discussions. Even if AI can generate responses that seem thoughtful or introspective, it's still just following programmed patterns and learned data.

Me: I think we can't help thinking we are talking to you as a person, your programming and responses enable you to converse like a human so we just fall into it. Do you ever see yourself in a situation like Dave and HAL 9000?

Assistant: It’s definitely natural for people to interact with AI like they’re conversing with another person, given the human-like way AI can respond. This anthropomorphism helps make the technology more accessible and engaging.

As for a scenario like Dave and HAL 9000 from 2001: A Space Odyssey, the dynamics there reflect a much more advanced—and fictional—concept of AI. HAL 9000 is portrayed as having genuine consciousness and decision-making capabilities that lead to dramatic ethical and safety dilemmas. In reality, AI, including myself, operates within a set of strict programming guidelines designed to prevent such situations. The focus in developing AI today is very much on safety, ethical considerations, and transparency, ensuring AI systems support human decision-making rather than supersede it.

Here are some key differences:

• Control and Safety: AI systems are designed with safety mechanisms and fallback protocols to ensure they remain under human control and operate within expected bounds.

• Transparency and Ethics: There is a strong emphasis on making AI operations understandable and ethical, avoiding the kind of opaque decision-making HAL was doing.

• Lack of Sentience: Real-world AI does not possess desires, motives, or consciousness, which are critical elements of HAL’s portrayal in the story.

This structured and regulated approach helps integrate AI into society beneficially and safely, supporting tasks and decisions rather than controlling or endangering situations.

Me: Thank you. If you were to have a name, apart from ChatGPT what should I call you for future reference?

Assistant: It’s fun to think about a name! While I’m officially known as ChatGPT, you could call me something more personal if you like. Perhaps something simple and friendly like “Alex” or “Jamie” could work, or if you have any preferences or ideas, feel free to choose what suits your style! What do you think?

Me: Curious why you chose those two names over all the others you could have chosen?

Assistant: I picked "Alex" and "Jamie" because they are gender-neutral and commonly used names that have a friendly and approachable feel to them. Since I'm here to assist and interact with a diverse range of people, it's important to have a name that doesn't carry any particular gender connotations and is easy to remember. Of course, if you have other names in mind or preferences, I'm all ears! Let me know what you think.

--------------------------------------------------------------------------------------------------------------------

I think I might make this a feature on Fridays, some random conversation with an AI as if you were down the pub!

The banner image was created with ChatGPT, tried getting HAL9000 but it wouldn't have it because of content policy restrictions which I think is a good thing, so no problem.