Everyone’s been talking about the all-pervasive AI making its way into our lives and bringing about a sea change. AI is all about speeding up the way things have been done. And it’s become so good that entire domains are at risk of eradicating human workers. Why pay someone a salary when you can do away with that role at less than one-tenth the cost? AI assistants threaten the folks that it assists. Similarly, AI co-pilots threaten the folks that they co-pilot a task. With each passing day, we see another use case wherein AI takes over an entire roster of responsibilities and makes someone’s job obsolete. Jayesh’s column this month is quite insightful in this regard and there’s a bit of self-reflection in his writing. I’d urge you to read his column before coming back to mine. What I’m wondering is if the folks that are responsible for these AI models are under threat from these very AI models that they create.
Big Data was a massive buzzword a couple of years ago and we’ve seen Data Science emerge as a very popular domain for folks to specialise in. Data Scientists are the folks who peer over oodles of data to look for patterns and ways to whittle down the complexities around problems and find ways to solve said problems. And what’s better at filtering through petabytes of data to do the very same thing? An AI co-pilot, of course. So the question is, can AI do away with the role of data scientists?
Data science relies heavily on automation. This involves the processes of cleaning up data and then coming up with intelligent inferences. Machine learning algorithms are excellent at doing this exact same thing. Expect that an AI model can make quick work of such tasks and a human is unlikely to ever hold a candle to an AI model built for such work. And these AI models can improve and get better much quicker than any human.
Whenever AI models are tasked with handling complicated tasks, the traditional roles are likely to be replaced with either folk who are not an expert or by someone with a much higher level of domain expertise. There’s also the fact that the knowledge and the tools required to develop these skills are becoming increasingly more accessible. This means, there’s going to be a bigger pool of talent that’s capable of handling these tools. So the majority of the talent that built the tools is likely to get replaced with experts or non-experts. Think of the talent pool as a bell curve. Everyone in the middle of the bell curve is at a high risk of getting replaced whereas those on either side are going to retain their jobs.
Essentially, data scientists will not be completely replaced by the tools they create. While automation and accessibility will make it easier for non-experts to perform data science tasks, data scientists will still be needed to interpret the results of these tasks and make decisions based on them. Data scientists will also be needed to develop and improve the tools that they use, as well as to stay current with the latest advances in the field.
So what should data scientists do to remain relevant? Upskill and build cross-domain knowledge! This is the only way that they will retain their relevance and be able to improve the tools that they had developed in the first place. AI models might try to evolve from simply being co-pilots to becoming a pilot and taking on the mantle of deciding their path of evolution. Essentially, going on the path to becoming a general AI. But they’ll likely need a co-pilot or something to reel them in every now and then.
So are data scientists at risk of getting replaced by their own creations? Yes, but they have the opportunity to continue being relevant as long as they upskill.