Creativity, Originality, and Intelligence
The debate about what it means to be 'human' is about to commence.
The stampede of people rushing to use ChatGPT for their business - such as content factories is almost evenly matched by those queuing up to find fault with it.
Here's the headline, in case you missed it. ChatGPT is not a maths tool, it's not a fact finding tool, it is a language generation and creative tool.
For the longest time, we we've been told that technology and automation can do a lot but can't take over from human creativity. This was always the last bastion - the ability to create. And generative AI tools have blown a hole through that argument.
Make no mistake this is an AI category that creates. Whether you ask it to create code, write a poem, paint a picture, or an essay.
We are now absolutely on the edge of the questions I'm sure the world will be asking soon. What is creativity? And in fact, what is to be human? Can you be creative without being human? The dictionary definition of creativity is "the use of imagination or original ideas to create something; inventiveness". Is generative AI using 'imagination'? That is definitely a human loaded word. What about 'original'? Bob Dylan says you're always singing into a tradition. Voltaire said originality is just judicious imitation, and that the most original writers borrowed from each other. So generative AI is doing its own kind of borrowing. It's finding it's own inspiration and masking it well.
Others have described originality as undetected plagiarism. Sholay is the arguably the biggest 'Bollywood' hit movie of all time, popular to this day, almost 50 years after it was released. The song 'Mehbooba' from the film is one of the most easily recognised songs in the pantheon of Bollywood songs. While Sholay was inspired by themes from movies such as the William Roberts' Magnificent Seven, itself a copy of Kuroshawa's Seven Samurai, the song Mehbooba is a much more direct lift from Demis Roussos' 'Say You Love Me'. Human creativity too, is therefore invariably diluted by inspiration and plagiarism. It's a particularly moot point as to where one ends and the other begins. Especially for a machine that is effectively just stringing words together based on a probabilistic model.
Whatever our ideas about creativity, this is not the first time technology has encroached into a creative profession. Two hundred years ago, photography appeared via contraptions called Daguerretype and the camera obscura. At the time painters tended to be realists, focusing on landscapes and portraits but mostly they painted exactly what they saw. Photography disrupted that model and over the next decades impressionism was born. Artists started interpreting reality partly as a reaction to having to compete with photography. And one of the questions I ponder often is about what might emerge as a response to today's technology encroaching into the territory of creative professionals.
I've had a number of discussions with friends about generative AI and ChatGPT. My friend Sriram said that it would make a big difference as to whether a response to a customer was generated by humans or machines. Which made me think - as a consumer - do we care if a reply from a service provider comes from a human or a machine? Or do we care more about whether it serves our needs better, and more effectively? What if the person behind the counter, or answering your call or message is just bored and doing a job?
If you went to see a movie, and enjoyed it, would it change your level of enjoyment if you learnt that the script was written by an AI? If you liked a piece of art, would it matter that it was not made by a human? After all, humans have been known to stick a banana on a wall, or put a urinal in a glass case, and call them both art. More functionally, would you care if your nicely worded response from the doctors' surgery asking if you're recovering from your last bout of illness, or from the charity you donate to telling you how your donations are being put to use, were written by a generative AI?
Today, generative AI can write essays, draw comic books, and have conversations. It can predict protein folding. And do a host of other things across the fields of arts, science, and engineering which are essentially creative acts. And it is still in its very infancy. A version 1.0 - the 256 colour, 8 bit version. As it grows in capability, which it will do very quickly, it will start to be more and more "human" or perhaps super-human. It will reach and surpass the uncanny valley stage. As our ability to combine generative AI with other technologies improves, such as this simple example of ChatGPT+Wolfram Alpha which can combine the language skills of the former with the computational capability of the latter.
The question of what it means to be intelligent is just around the corner. For thousands of years, we thought that the sun and all other celestial bodies revolved around the earth. Despite many hypotheses to the contrary, from ancient Greece, India and other places, it was only after Copernicus and Galileo that we came to believe in heliocentricity. This also shifted our philosophical plane and our appreciation of our position in the universe. A similar discussion may be about to commence on our current anthropocentric worldview versus a view where humans are just one of many intelligent forms. After all, in a different turn of evolutionary events, it might have been the Octopus that was the dominant conscious life form on earth. We are one of many intelligences, and going forward we might not be the smartest, in many ways.
More Stories About Generative AI
Azeem Azhar on the ChatGPT business model (LinkedIn)
Microsoft and OpenAI’s investment model (Wired)
ChatGPT’s impact on the classroom (Wired)
GPT’s Protein model performs impressively (IEEE Spectrum)
John Battelle on why Conversational AI will make us Centaurs (Medium)
Reading List
Platforms: Cory Doctorow’s excellent essay on why platforms tend to decay in terms of their usability and value to consumers a process he calls enshittification. TL:DR - they start with serving users, then they move to serving suppliers/ advertisers, and finally cash out by serving investors. (Wired)
Food: I’ve been very bullish about plant based meat and continue to believe we’re seeing a big shift. This piece argues otherwise, calling out some health concerns and technical challenges. Feels a bit like a hatchet job, though. We still make Beyond Burgers at least once a month at home. (Bloomberg)
Product Innovation: A peek into the innovation environment at Converse Labs where the Chucks shoes take on a lot of cultural and environmental territory. (Fast Company)
Magical Science: This phase changing robot is very clearly in the category of tech that is so advances so as to appear magical. This is phase changing metal. Can you imagine the applications of this?? (YouTube)
Economic Bubbles, A great piece on Charles Mackay and the railway bubble which drew in amongst others the Bronte Sisters and many more, by Tim Harford.
Climate Change: What will it take to get past fossil fuels? A closer look. (Economist)
Climate Change: A great blending of design and technology - these Tera Cotta tiles also double up as Solar panels (Fast Company)
Tech futures: what happens when technology becomes invisible or even boring? Answer - they become really useful. Or perhaps, they become invisible when they become useful! (Medium)
Tech Futures: A spray on ‘skin’ that is made of smart materials which can recognise your gestures? Why, yes, right this way. (IEEE Spectrum)
Gender balance: Could the gender balance be shifting? This piece might sound like a cry for help for boys, but really the more positive aspect of it is that there might be signs that women (girls in schools and colleges) are making up ground and the future of work might have a very different gender balance. Don’t hold your breath - there’s a long way to go. (New Yorker)
Retail: is the high street on its way out? (FT)
Retail: is this retail’s next big challenge? How to crack the live stream shopping market? It’s those influencers again! (FT)
Electric Vehicles: EVs crossed 10% of all new cars sold last year- a critical milestone. (WSJ)