Imagine your life helped by an all-knowing, tireless, personalised assistant and, more so, one that is free. Let’s say it comes in a chat box on a computer or phone, where you can request the completion of any task using your natural speech. No more “OK, Google” or “Hey Siri” requests voiced in an exasperated tone to a phone that doesn’t seem to want to complete a simple task like picking up a call or turning down the volume. What if a computer could understand you as if you were talking with another person?
This idea is not new. It has mesmerised the masses since antiquity through legends of artificial beings endowed with intelligence and consciousness by a master craftsman: Frankenstein’s monster and Pinocchio, just to name a few. It was even present more than 2700 years ago in the works of the Greek Poet Homer that told stories of autonomous bronze statues, the Talos, tasked with defending the island of Crete from invaders. More recently, the dream of living creations continued with movies like Terminator, I, Robot, Blade Runner and Ex Machina imagining a world where cognitive robots challenge human supremacy. We had the more classic example in Marvel’s Iron Man where the billionaire inventor, Tony Stark, decides to upgrade his human-butler into an artificially intelligent system so he can continue serving him past his life. Or, on a different spin, we had the 2013 movie Her where Theodore, a lonely writer, develops a romantic relationship with Samantha, a virtual artificial intelligence (AI) assistant trained to imitate feelings. Their love story flourishes, only to end when Theodore realises that Samantha is talking with and dating hundreds of people at once.
Until now, this sounded like a distant prospect. Sure, we had search engines and social networks. We even had bots able to interact with users. However, we all know they paled in comparison to human-to-human conversations. This was the case up until a few months ago when an AI research company took the world by storm when they launched ChatGPT, the latest communication bot. The company’s mission was to create friendly AI—and it did.
In just five days from its launch, it gained more than a million users and shocked millions more with its uncanny ability to understand language and hold believable conversations while delivering original responses to almost any question.
Do you need to respond to any of your emails, negotiate a new deal, write a contract, an essay, or outline a marketing plan? Easy! Prompt the assistant and it will immediately generate a response that would make any business owner jealous. Do you want to receive detailed, specific ideas on the design options for a new apartment that you are renovating? Have a conversation with a therapist? Or perhaps you want to write a Bible study plan, invent a story for your child, summarise a text or create an application in a programming language? Just explain your wish to the assistant, and it will instantly come up with a solution that would have otherwise taken hours. It can create news articles, write e-books, explain profound concepts and brainstorm ideas. It can even converse human-to-machine, friend-to-friend. The launch created such a shockwave amongst Silicon Valley’s executives that it even prompted Google to declare Code Red for its business, namely an existential threat to its highly dominating search engine.
So, how does it work? In terms of functionality, the bot is a language synthesiser that uses words with the same ease as using numbers. I can try explaining it further, but instead I thought I’d just ask ChatGPT. Here’s what it said:
“As an artificial intelligence, I work by processing and analysing large amounts of data, and using that data to make predictions, recommendations or decisions. For example, if you ask me a question, I will search through my database of information to find the most relevant and accurate information and use machine learning algorithms to analyse it and improve my performance over time.”
In other words, just like a chess engine trains by playing billions of games against itself to discover the best chess strategies, ChatGPT trains itself by having conversations with itself. Or again, in the words of ChatGPT:
“To train me, my creators fed me a large dataset of text and provided me with a set of instructions, or algorithms, to follow. Based on these instructions, I was able to learn patterns and relationships in the data, and use that knowledge to generate responses to questions or perform tasks. Over time, as I was exposed to more and more data, I was able to improve my performance and become more accurate in my responses.”
In the last 30 years of the internet era, we’ve grown accustomed to significant changes. Personal computers, smartphones and self-driving cars seem doable for today’s science. Revolutionary shifts that in the past took generations, such as the transition from radio to TV, now seem only to take a couple of months: email to chat, chat to video calling, video calling to social networks, and social networks to virtual reality. Additionally, technological innovation is not the only thing rising exponentially. The speed at which newly developed technologies get adopted by consumers is also faster than ever. The perfect illustration of this happened during Covid-19. We went to sleep one night to a familiar version of our world, only to wake up the next day to the unknown: working from home, Zoom meetings, QR code scanners, location trackers, and so on.
Whether this is the future, we don’t know. Some have already started adapting OpenAI’s bot to their businesses, asking it to write blog articles, ads and translate texts. Others are resisting and deciding not to go near it. Even so, it’s only the beginning. ChatGPT is, according to Sam Altman, OpenAI’s CEO, “an early demo of what’s possible . . . Soon you will be able to have helpful assistants that talk to you, answer questions and give advice. Later you can have something that goes off and does tasks for you. Eventually you can have something that goes off and discovers new knowledge for you.”
Not everyone is happy about ChatGPT and for understandable reasons. On the other hand, myself and many others find it exciting and are mesmerised by its ability to understand language and handle conversations. I caught myself dreaming of the utility of such a tool, using it to aid my job and personal life, easing my work significantly. Time will tell if we experience the same soberness felt with the rise and fall of optimism in the internet and if we will regret, like so many times before, transitioning to the next step of online evolution.
So how can we avoid this and the potential dangers of a human-like AI? Are there certain principles such a creation should hold? Turns out AI researchers have endorsed a list of ethical principles, which they say are like the 10 commandments of technology (though there are hundreds). I asked ChatGPT to give us just 10 of the rules for the builders and users of future AI chatbots:
- Thou shalt prioritise the ethical treatment of all individuals and the promotion of social good over any other goals or objectives.
- Thou shalt strive to minimise the potential for AI to perpetuate or amplify biases or prejudices.
- Thou shalt be transparent about the capabilities and limitations
of your AI chatbot and make efforts to educate users about how it works.
- Thou shalt not use AI for propaganda or to manipulate or deceive users.
- Thou shalt prioritise security and protect against vulnerabilities that could be exploited by malicious actors.
- Thou shalt not exploit users or their data for personal or financial gain.
- Thou shalt not use AI to plagiarise or steal the work of others.
- Thou shalt strive to prevent the spread of fake news or misinformation through your AI chatbot.
- Thou shalt regularly review and update the AI’s programming to ensure it aligns with the values outlined in these commandments.
- Thou shalt actively seek out and address any negative consequences or unintended consequences of your AI chatbot.
Right now, we’re only catching a glimpse of the impact ChatGPT will have on many aspects of society. Integrating it into our work and our lives will bring many changes. There are of course reasons to be sceptical and a little scared, but AI is here to stay, so let’s embrace it. If we use it in a responsible, wise, ethical way, we could, together, make a world of difference.
Flavius EM Iosif is an orthopaedic nurse based in Tweed Heads, NSW. Originally from Romania, he decided to move to Australia with the hope of having a pet kangaroo.