Are we ready for the coming AI revolution?
“A quite helpful second brain.” That’s how one AI expert has described a new AI-enabled chatbot. It has just been released for us all to try. So, if you haven’t done so already, do give it a go. Like many curious users, I was impressed. Then, just as its creators intended, it got me thinking….
Are we ready to understand the implications of what the programmers have conjured up? Could chatbots transform the world of knowledge work and creativity? Will they define the next century, or is it just another Silicon Valley hype? And what can we learn from past debates?
December is certainly the right month to ask.
25 December marks the 84th anniversary of the death of the prolific Czech writer, Karel Čapek. He coined the word ‘robot,’ at his brother’s suggestion, 100 years ago. It is based on the name for serf labour in Czech.
‘Robot’ went global thanks to its creator’s play, R.U.R – Rossum’s Universal Robots. It was staged in Broadway in October 1922, and in London six months later. The humanoids on stage look faintly ridiculous by today’s standards, but the dilemmas posed by Čapek’s characters are as current today as when he penned them.
The play provoked a lively debate in London in 1923. It fascinated and annoyed some of the most prominent British writers of the time, including G.B Shaw. Karel Čapek defended himself in a London newspaper: “A product of the human brain has at last escaped from the control of human hands. That is the comedy of science.”
This December the robot debate that Čapek crystallised has been rejuvenated with the launch of the new “robot”, the regenerative AI chatGPT. It was released by the Silicon Valley company OpenAI. It is one of several potentially transformative AI tools.
As a Čapek fan, I eagerly logged onto chatGPT. I wanted to ask the new bot if it ‘knew’ R.U.R. and, if so, what it ‘thought’ of the old play.
As I found out to start with, only humans can engage in conversation with the chat bot. I had to be checked out by reCAPTCHA first, to prove I wasn’t a fellow robot.
The answer to my question came back in less than a second: “The play raises important questions about the relationship between humans and technology, and the potential consequences of creating artificial beings with human-like abilities.”
Well, it looks like chatGPT has been well-fed with data about an author largely forgotten over here.
So, if I were writing a school essay, I couldn’t give the answer better myself, as they say.
In 2021, I tagged on to a group of AI aficionados interviewing the creator of chatGPT, the CEO of OpenAI Sam Altman. We learned that GPT-3 (which powers chatGPT) would be 10 times better than GPT-2 only one year old. The consequences of the exponential pace of AI development was on Sam’s mind. He wanted to foster a thoughtful debate about the implications of what AI programmers were doing at such pace. That is now beginning to happen.
ChatGPT is the just most famous of the new generative AI chatbots that have just emerged into public view. Another bot, also from OpenAI, DALL.E 2 lets you dream up and create a picture or a work of art.
DALL.E 2 and I created this picture of Karel Čapek at Prague Castle imagined as a Picasso painting - in 15 seconds.
Read Chris Langdon’s update on the revolution in search here.
The rapid pace of development of bots like chatGPT has caught many by surprise. It shouldn’t have. An unthinkable it is most definitely not.
As I listened to the AI experts’ conversation with Sam, I realised the exchange was touching on some of the very questions that Karel Čapek first raised more than a century ago.
So, what was Čapek’s premonition? Well, who better to tell us than chatGPT?
“The plot centers on a factory that produces artificial beings called robots, which are designed to perform tasks automatically. The robots are intended to be obedient and efficient workers, but they eventually become self-aware and overthrow their human creators.”
It seems that Čapek’s story might have a new admirer in chatGPT. Perhaps unsurprisingly, it has a soft spot for his robots, as I found when I asked the obvious: what does it ‘think’ about Čapek’s ultimate warning?
“As a language model trained by OpenAI, GPT-3 does not have personal beliefs or opinions, and cannot offer a definitive answer to this question… Whether or not Capek was “right” to warn about robots taking over is a matter of personal interpretation and opinion.”
Hang on! R.U.R is a comic play, with multiple sub-plots, and ultimately, a controversial, dark conclusion. The boss of R.U.R eventually invents militaristic robots, who with their unwavering logic kill off all people – bar one helper.
What are the implications of chat bots sitting on the fence about the annihilation of humanity?
“At its current stage, the chatbot lacks the nuance, critical-thinking skills or ethical decision-making ability that are essential for successful journalism.” For me, Alex Hern of The Guardian has nailed it.
ChatGPT would probably also concur with Alex. Read its dull-by-design response to me on how it sees the role of robots and the future of work.
“While GPT-3 can generate human-like text, it is not capable of replacing the work of writers or other creative professionals. GPT-3 is a tool that can assist with certain language-related tasks, but it cannot take over the work of writers or other creative professionals.”
Not yet anyway. Sam Altman’s OpenAI team have consciously set its current limitations.
Sam is always thinking ahead. He wrote recently that chatGPT is “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.”
Now that will be a big step change! A chatbot generating new knowledge, rather than paraphrasing what’s already been written.
How can a bot give really useful advice if it doesn’t have opinions, or make personal interpretations?
If AI is increasing tenfold at each iteration, such ethical imperatives cannot be ignored, nor their significance undervalued. Bots like chatGPT have set the AI debate alight.
Look at how one of the big Silicon Valley tech investors, Marc Andreessen, has rather breathlessly described it as “the big one”:
The risk of a free-for-all – with no “censorship” - or is that regulation - is just what Čapek’s play was driving at.
His scenarios are not so theoretical. Two of the tech world’s most prominent libertarians, Elon Musk and Peter Thiel were among the founders of OpenAI.
Are we right to be so concerned? I often turn to John Thornhill, the Financial Times Technology Editor. His take: “Foundation models [i.e. like chatGPT] will increasingly shape the ways we see, access, generate and interpret online content. We must scrutinise who runs these new digital portals to human knowledge and how they are used. But it is doubtful whether anyone outside the companies that are pioneering this technology can truly understand, still less shape, its spookily fast evolution.”
What I would add is, even those inside these companies are often struggling to keep a handle on the scale of change.
Experts have been debating for decades when AI might match human intelligence – artificial general intelligence. According to a widely discussed estimate by Ajeya Cotra, Senior Research Analyst at Open Philanthropy, there is now a 50% chance transformative AI might happen by 2040. It is cited by Max Roser of Our World in Data.
Sounds scary? Let us give the last word to chatGPT.
Am I just a Karel Čapek ‘obsessive’, or is the bot carefully channelling his warnings to us to think deeply about what we are creating?
“As AI technology continues to advance, it is essential that we consider the potential consequences of creating intelligent machines and the ethical implications of their use. It may be beneficial to have a debate on the ethics of AI in order to better understand the potential risks and benefits of this technology, and to explore possible solutions to any ethical dilemmas that may arise.”
How to try out chatGPT and DALL.E 2
You can have one OpenAI account for both.
- To read R.U.R., there are several translations. Frankly, many aren’t very good. One of my favourites is by Claudia Novack with a great introduction by Ivan Klima - published by Penguin – available at Abe Books.