World, Meet AI
You’ve been using it all along...
Do you have any idea how advanced you are in using technology? Yes, you with the beaten-up old Samsung S3 and you with the latest iPhone. You computerphobe and you early adopter. And everyone else in between. If you use a smartphone, you use AI.
The moment you turn to Google Maps, Apple Map or Waze to give you directions and help you avoid heavy traffic, you are using AI. When you use predictive text in WhatsApp and iMessage, and grammar correction tools in Gmail and Word, you are using AI.
‘The first secret of AI is not how it works or how it will change the world but rather how it is already embedded in so many things we do every day, and in so many tools we take for granted.’
In truth, AI has been practising on you, preparing for the big game when it comes out to put its skills on display in ways we never imagined possible. It’s not for nothing that the creators of new chat tools for generating content, images and videos describe the process of building these bots as ‘training’, whether on large language sets or vast image libraries.
Beyond these very publicly self-acclaimed AI revolutions, AI has been in training for many years, and that training is still going on. And you are an unwitting but willing tool in that training. If it can make your life easier, why wouldn’t you let it? With this in mind, it is clear that what we are seeing today are the demo models, early versions of mainstream AI tools that give us a broad idea of what they can do. When they are fully trained and integrate learnings from the training of numerous other AI tools and models, they will transform any activity one can imagine.
That is the second secret of AI: it is able to draw on all of recorded human knowledge to deliver anything, from an answer to an article to an archive of wisdom. To anyone. Anywhere. At any time. It could even have produced the alliteration in those last sentences on demand, although in this case it did not. I promise. AI will make us all superhuman if we want it.
Definition of AI
Scientists and software engineers tend to define AI in arcane terms, but the cat is out of the bag. I asked each of the three major generative (systems that can generate content on demand) AI platforms to give me a definition in one sentence for the layperson. To be exact, I said: ‘Please give me a one-sentence definition of artificial intelligence for the ordinary, non-technical reader.’ They produced this:
OpenAI ChatGPT: Artificial intelligence refers to computer systems that can perform tasks that typically require human intelligence, such as learning, problem-solving, and decision-making.
Google Bard: Artificial intelligence (AI) is a branch of computer science that deals with the creation of intelligent agents, which are systems that can reason, learn, and act autonomously.
Microsoft Bing Chat: Artificial intelligence (AI) is the ability of a computer or a robot controlled by a computer to do tasks that are usually done by humans because they require human intelligence and discernment.
Yes, I also found that dreadfully dull. And that is the third secret of AI: if you have access to all the knowledge and wisdom in the world and are asked to bring a complex subject down to one sentence or paragraph, you will average out all definitions available to you and end up with fundamentally average content.
The question I asked is known as a prompt. The quality of your prompt will also partly dictate the quality of the answer, and an average prompt will produce an average answer.
So, I asked the question again, but this time I asked each of the platforms to ‘make it humorous’. This is what they told me:
OpenAI ChatGPT: Artificial intelligence is like having a computer buddy who is so smart, it’s like they went to Hogwarts for coding.
Google Bard: Artificial intelligence is when a computer can do something that you’re pretty sure you could do if you had a few more hours and a lot less coffee.
Microsoft Bing Chat: Artificial intelligence is like a five-year-old who can do calculus.
I confess I was surprised how quickly they caught on. That last one is even subtly self-deprecating, without being asked to be.
What this tells us is that AI is undergoing a fundamental shift, from a focus on what it is programmed and designed to do, as with bees and football leagues, to a focus on what we ask it to do. This is the fourth secret of AI: just as the quality of programming is key to the underlying power of AI, so is the quality of our input the key to what it can produce.