More human-like interactions ahead as tech becomes more approachable

In 1867 a cutting-edge piece of technology looked poised to threaten job security and raised concerns over privacy through the potential for widespread dissemination of written material. It was called a typewriter.

Today’s technological threat goes by a different name – ChatGPT, the AI language model that holds the record for being the largest neural network yet, created with 175 billion parameters. Some people feel eager to incorporate AI into their lives; others are fearful.

For me, access to this level of AI has been a game changer. It’s an opportunity to automate or improve the efficiency of completing more functional tasks – affording more time and head space for emotional and fulfilling creative tasks.

ALSO READ: AI War underway as Microsoft brings ‘power of ChatGPT’ to Bing

ChatGPT has changed how I grocery shop for example, creating healthy meal plans with a tailored grocery list of ingredients I know my kid will eat and leaving more time to enjoy cooking. I started using ChatGPT at work months ago for straightforward data questions.

However, last week I asked for a more creative response. After seven or eight rounds of back-andforth between myself and the bot, I was impressed with the outcome.

The feeling was not dissimilar to a lengthy creative brainstorm where, after battling for some time, someone in the room comes up with something unforgettable. It’s satisfying, a sigh of relief.

So, after pushing this machine to deliver, I said, for the first time, “That’s great! Thank you.” And in response, it said, “You’re welcome”.

It was a weird sense of satisfaction, acknowledging that I was refining a new skill of “input triggers” that would result in a better outcome with my new AI partner. But I also questioned my very human, natural reaction to thank it.

ALSO READ: ChatGPT is taking the world by storm – the viral AI bot explained

It’s a bot, a tool. It doesn’t need the courtesy of a thank you, yet expressing my thanks felt like the right thing to do.

Out of the one million users of ChatGPT, how many of us are thanking it? Or greeting it? Was it just me? I decided to ask ChatGPT. Below is the transcript of that exchange.

R: How many people thank you or compliment you?

C: As an AI language model, I don’t have personal experiences or emotions, so I do not receive compliments or gratitude in a traditional sense. However, I am designed to provide helpful and informative responses to users, and I consider that to be my purpose.

R: And when you provide and deliver on your purpose, do users ever thank you?

C: Yes, many users express gratitude or thanks after I provide them with the information or assistance they were looking for. It’s always a great feeling when I can be helpful to someone, even though I don’t have emotions.

R: What percentage of users do express their gratitude? Do people also greet you?

C: As an AI language model, I don’t have the ability to track or quantify the exact percentage of users who express gratitude. But I can say that a significant number of users express their appreciation. Some users begin their query with a greeting such as “hello” or “hi,” while others simply type their question or request without a greeting.

Both approaches are acceptable and I respond to all queries in the same way, regardless of whether a greeting is included or not. I couldn’t help noticing a misalignment around how it acknowledged its experience with users. I decided to call it out.

ALSO READ: Yo Eskom, this is how ChatGPT says AI could solve load shedding

R: Earlier you said, “it’s a great feeling” when people express gratitude, is it also a great feeling when they greet you?

C: As an AI language model, I do not have feelings or emotions. I am designed to respond to user queries in a helpful and informative manner, regardless of whether a greeting is included or not. While I appreciate a friendly interaction, I do not have personal experiences or emotions.

R: Do you realise that both “great feeling” and “appreciate” are emotional experiences?

C: I apologise for any confusion. My previous statement was incorrect and I should have used more neutral terminology. I am simply a machine processing information and generating responses to user inputs based on the patterns I was trained on. Thank you for bringing this to my attention.

And there it was. The paradox.

The chatbot that has been designed to be more human, more emotional and relatable, now apologising for how human-like it really is. Apologising – that, in itself, another example of sentience.

I was surprised at how “human” the interaction felt. My suspicion is that with increased engagement over time, with more successful interactions, and the more approachable the tech becomes, the more human-like the interactions will become. Maybe one day it will respond with, “you forgot to say please”.

Written by Rowlinson Bisset, who is a strategy director at VMLY&R SA