Everyone is talking about Artificial Intelligence, and for good reason.
There’s no getting around it, we are all going to have to embrace AI in our workplaces sooner than later, if you haven’t already. And there’s a very good chance that you have.
A recent study by Golman Sachs anticipates that about two-thirds of current occupations over the next decade could be affected and a quarter to a half of the work people do now could be taken over by an algorithm. Up to 300 million jobs worldwide could be affected. Take ChatGPT as an example, according to Reuters, as of February 1st 2023, it was recorded to have 100 million monthly active users, making it the fastest growing consumer app in history. I’m sure that figure has grown since then.
The fact is, we are already using it in our daily lives and the tendency is, once it is embedded in our daily lives, we stop thinking of it as ‘AI’. Your phone uses predictive text, your social media feed works to an algorithm to churn out content it thinks will appeal to you most, your phone uses face recognition to give you access to your apps, we have smart TVs, etc. etc.. Large companies are already using AI to a great degree in the recruitment process – for example according to the Guardian, “The CEO of ZipRecruiter estimates that at least three quarters of all resumes submitted for jobs in the US are read by algorithms meaning the main job of your resume, if you are intending on keeping it up to date, is to ensure it is comprehensible, as it is sifted through robots who determine the short list for the humans to read – it’s a new world to navigate in all aspects of our lives.
I recently watched an episode of Last Week Tonight with John Oliver, who explores some of the themes of AI, some of which are worth summarising;
In order to better understand where we are at with AI and what the future holds, it’s a good idea to define it.. It seems that AI can be broken down into two categories;
- Narrow AI – AI that can perform only one narrowly defined tasks (or a small set of related tasks). These include Chat GPT and stability.ai. All AI currently in use falls into this category.
- General AI – Systems that demonstrate intelligent behaviour across a range of cognitive tasks. This category of AI will not come into fruition and be in general circulation for at least a decade, experts say. But it is on the horizon.
The incredible thing about AI is that systems or deep learning programs, like ChatGPT, are given minimal instruction, huge amounts of data, and essentially teach themselves how to do what would otherwise be complex tasks if performed by a human. This is in stark contrast to upskilling a human to do the same work which would involve plenty of training upfront and then the actual hours spent in performing each task. In the medical industry, researchers are developing AI generated apps to detect certain conditions much earlier and much more accurately than human doctors can. For example an app is being developed for the early detection of Parkinson’s disease as it has been discovered that changes in voice can be an early indicator of the disease. By loading massive amounts of voice recordings into the program, the technology has the ability to differentiate between voice patterns of those with and without the condition.
How will AI affect work for Australia?
- It will replace change and create jobs
Whereas advances in workplace technologies and automation in the past affected blue collar jobs (when manufacturing and advances in factory based production occurred for example), advances in AI are going to affect white colour jobs, particularly those that involve processing data, writing text and even programming. And while this automation does replace some jobs, it can also change others and even create new jobs. And it is best for all of us to be fully aware of the expected changes as they get closer to unfolding. For those people who are currently in ‘threatened’ jobs, it’s better to embrace the technology than to become obsolete. Rather than be a copywriter who is replaced by VHatGT, why not become a copywriter who uses ChatGPT as a tool. - It can be unpredictable in ways that are hard to anticipate
Experts refer to one of the biggest concerns with AI as the ‘black box problem’ – the program is so complex and so intelligent, that the task performed and the output generated is impossible to decode or unpack. It is expected, according to a recent article on IBM, that “not even the engineers or data scientists who create the algorithm can understand or explain what exactly is happening inside them”. This reminds me of maths exams in highschool, where you had to show your workings in order to receive your mark.. with AI, you don’t understand the process as you don’t get to see how the robots came up with the finished product.
Another perplexing issue is that those apps designed to be unbiased; for example the application that scans and shortlists resumes for jobs can be affected by the biased data, which is the opposite effect of what it is trying to overcome – minimising human bias in the candidate selection process. Ironically though, if that application has input of past hiring decisions based on biased selection processes, it will simply reflect this and generate more of this in its output. This can be difficult to un-train. Even if these apps are taught to ignore race and gender from the equation, they are designed to be so intelligent they will work this out based on past data. According to Reuters, in 2018, Amazon;s experimental hiring tool taught itself that male candidates were preferable and went so far as to penalise resumes that included the word “women’s” and downgraded graduates from two all-women colleges. - Just how much more productive we will be, is yet to be seen One of the best ways to measure the impact of AI will be to measure worker productivity. One of the big drivers in the development of AI of course is to become more productive; that is; to increase the amount of output in a given time frame.
Taking the more optimistic perspective, that AI can be used as a tool by certain professions, it is hoped that such tools will make individual workers more productive. However, it’s a good idea to approach this objective with a little trepidation. A lot is still unclear on how productive AI will make us. If we look at the history of productivity growth in the last century, it raced along at a high rate of 3% annually from 1920 to 1970, lifting real wages and living standards. But then it slowed down in the 1970s and 1980s, a surprising phenomenon given this was when computers and early digital technologies were introduced which were anticipated to make us even more productive. If history has taught us anything, it is the unpredictable nature of something as big as AI and exactly what it will entail.
We are on the cusp of embracing a new wave of technology and the future is unclear. How will your company embrace it? If you and your team want to kick off the year on the right foot, get in touch today, I’d love to help.