I hope this email finds you well and that your January has gotten off to a rocking start. We’re flying at the moment.
This is a testament to a lot of businesses looking to make decisive moves early in the year – something that we wholeheartedly encourage! (Frankly, it’s great to gain forward momentum when others are lagging at the starting block!)
In this week’s newsletter, I want to talk to you all about AI and machine learning and ask a pretty thought-provoking (and somewhat scary!) question: “could a robot do your job?”
Background
Before we get into the controversial stuff, let’s do a quick crash course in artificial intelligence and the state of play – so we’re all on the same page.
AI is a term widely used for machine learning technologies. These essentially apply a lot of computing power to review large data sets and output text, images and all sorts of other things based on a request from a human operator.
It might be a language model which produces a poem about spaghetti in the style of Lord Byron or a picture of Peter Kay in the style of Rembrandt.
Over time the outputs of machine learning technologies have become increasingly accurate and compelling. What’s more, the better it gets, the better it gets at getting better (if you catch my meaning)… This is resulting in an exponential year-by-year improvement for AI.
Crucially, as machines in the past have removed the need for physical labour, so AI promises (or indeed, threatens) to displace cognitive labour, potentially being disruptive to the way we work and interact with the world more generally. It genuinely has the potential to be the equivalent of machines that replaced a lot of physical labour from the beginning of the 20th century during the industrial revolution.
Computers that learn are not a new thing, of course, as we all know that given the behaviour we input into social media, the internet and YouTube. These sites have become incredibly good at learning about us and presenting us with more things that it thinks we might find engaging (if any of you have tried TikTok, the extent to which it can present you with something novel at the flick of a finger is frankly scary!)
The ”AI revolution” has been something that has been on the horizon for a long time, but with the arrival and availability of “ChatGPT” to the public at the end of 2022 – for many, it seemed like that moment had been reached.
Let me explain.
Then along came ChatGPT
So, what is ChatGPT? ChatGPT is a language-based AI model from OpenAI, an AI research company based in San Francisco, California. OpenAI has received $1 billion in investment from Microsoft, and it is purported that Microsoft will look to increase their funding of the company to the tune of $ 10 billion – and potentially include the AI in their Bing and Microsoft Office products.
But how does it work exactly?
Well, the model is trained on an incredibly large dataset – with all sorts of information leading up until 2021 (which is why it doesn’t know about new things). What it does when it is asked a question is predict what the most likely and accurate response is on a token-by-token basis.
For the user, it feels like you’re chatting with another person, who is incredibly smart and can output text about complicated topics very quickly and with a high degree of complexity.
And it works! If you ask it a question it comes out with a very compelling, in-depth, and typically accurate response. (Though it should be noted that it can be wrong and in those instances, it would be very confident despite being factually incorrect because of how it is programmed.)
[Incidentally, a lot of people are comparing ChatGPT to Google and saying that the richness of the answers it provides is a threat to the search giant. Despite this, the technologies are fundamentally different. Search engines index and present the most relevant information based on the search query, whereas a language model like ChatGPT composes its response “from scratch” based on the most likely and accurate next “token”, or phrase that should follow. It will be interesting to see how this might be integrated into Bing to improve its search – it may increase their search share as people find ChatGPT responses more natural and in-depth.]
Crucially, you can ask it to do things like: “give me a top 10 list of innovation methods for small businesses” or “write a content outline for a blog about IP and international trade” and it will be able to do a LOT of the heavy lifting for you – and that is it in its current form.
This year a new version is going to be released that is said to eclipse last year’s capabilities and so it will be for all future versions released. It’s also important to note that ChatGPT is also not the only AI in development, with every large entity investing in and developing AI in one form or another.
ChatGPT and AI at large and the impact it will have on everything are here, and it's not going anywhere.
Ultimately, along with other ever-improving models that can produce high-quality images, 3D models, videos and more that are indistinguishable or “good enough” for deployment in all sorts of different places – including, dun, dun, dun – the world of work. And the dissenting voices who fear that it may begin to replace certain people’s jobs are not entirely incorrect...
The robots are coming, what do I do!?
Whilst certain industries are already seeing disruptions as AI starts to produce outputs that displace or replace work that would be undertaken by a human (copywriting and illustration, for example), it won’t stop there.
Put simply, we’re entering a world where nearly any cognitive task may be better undertaken by AI like ChatGPT or its successors in the first instance, and it is something we must be ready for.
It is equally applicable if you’re writing a social media post or a legal case write-up.
Frankly, it’s scary, and it is something that is moving faster than many people will be ready for.
There will come a time when AI helps us produce most of the output. In as little as 20 years, I believe it will be seen as far less efficient to not do this and seen as very old fashioned!
So, if we end up outsourcing our initial thinking to robots where does that leave us all?
Crucially, AI only seems to give as good of an answer as the questions it is given.
As such, the role of the worker may become less of a creator of information but of a curator and editor. You see this with programmers outputting code – you have to know what you’re talking about to get anything usable. The same goes for law.
You could ask an AI to write a trade mark opposition for you – but, how would you know it fulfils your commercial aims unless you know what you’re looking at when it outputs?
Ultimately, tools such as these which stand to empower each and every one of us can be scary – but they are equally exciting. I for one am hotly anticipating the boom in innovation and productivity that may well result from the wider use and adoption of ChatGPT and its likes.
So, can an AI do your job?
Not yet, probably.
Soon, maybe.
But there are some things that it will never be able to do, and I think a focus maximising our time on that is where those who grab the possibilities offered by AI by the horns will really make an impact!
Wrapping up
But what do you think?
Will AI take all our jobs and leave us with nothing to do; will it supercharge you and your employees to achieve far more than you are now, or is it all a storm in a teacup?
I would be interested to hear your thoughts!
Also, as a parting question…
If AI could do one thing for you in your role, or business - what would it be?
I would be very intrigued to hear from you about that.
Hope this newsletter finds you well and that you’re having a brilliant week.
As always, if you want to have a chat about anything, drop me a line, or click the button below to shoot me an email. My door is always open.
