AI won’t replace people, yet. But people who use AI will edge out those who don’t

There are clearly those that see so much potential in the technology, that human roles could be rendered redundant due to AI.

Ever since the release of generative models to the public, there has been a lot of debate and anxiety about whether we need to worry about our job security. And, rightfully or wrongfully, many alarmists have been warning us of a dire future full of unemployed masses.

However, there is a lot of nuance to this debate, based on how you slice and dice the economy. There’s government vs private, organised vs unorganised, established firms vs startups, large companies vs small enterprises, and many more permutations that will need addressing, before any final call can be made on the future of employment in the age of AI.

This article, in support of nuance, is just an attempt to explore these various dimensions of the jobs-enigma.

First, we look at the optimists. Recently, at a panel discussion at the B20 Summit, N Chandrasekharan, Chairman of Tata Sons, stated that AI will lead to the creation of various new jobs. This will happen, he said, as people with little or no skills will be enabled by AI to become capable of performing higher level jobs.

Similarly, Accenture, one of the largest absorbers of fresh STEM graduates in India, has stated its intent of creating 40,000 new AI jobs to provide services across 19 sectors, and has made many acquisitions to support its vision. It plans to spend over $3 billion over three years in its data and AI practice.

Thus, powerful players, who are in the business of providing mass employment, are looking at the emergent technology of AI as a new tool that will only add to their workflow but will also create more possibilities.

Challenging the above notions, firms such as IBM have explicitly stated their intent to replace roles with AI, with the organisation quoting an 8,000-figure mark to show the scale of its intent in replacing humans.

Further, recent data at an industry-wide scale in the USA indicates that of the 80,000 layoffs that occurred in May this year, nearly 4,000 were due to AI. Thus, there are clearly those that see so much potential in the technology, that human roles could be rendered redundant due to AI.

So, what’s the middle ground here? For a nuanced answer to this question, we can look at various surveys to see how professionals themselves are feeling about their job risk due to AI.

And, in one poll after another, the trends seem to be non-alarmist. While most professionals appreciate the capabilities that AI is providing in improving communication and automating facets of their work, they don’t see how such systems could come to replace the human element wholesale.

Moreover, reports on AI adoption, and its poor take-off compared to expectations, shows that most organisations are still playing around with the technology and imagining use cases where they could leverage this technology.

So, going ahead, for any discussion on jobs, we should look at the underlying dynamic and technology of AI, and extrapolate from there. The first thing to keep in mind here is that AI is at a nascent stage of development with its competencies extending to those of a capable assistant rather than a coworker or boss.

This can be understood by just looking at the rates that AI bellwethers are charging clients for their tools, such as Microsoft’s fee of $30 per person per month. Had they been confident in their ability to replace workers wholesale, the prices that were being charged would have been much further north of the figure that they have. Thus, AI, at its current level of competence, will not lead to mass unemployment.

Second, there are only a few players in the market that can make the chips needed at the scale required for making the next leaps in generative AI. This can be inferred from Nvidia’s meteoric share price rise, which has allowed them to break into the trillion-dollar club.

From ideation to design to fabrication to deployment to training, there is a long list of things to be accomplished before much larger scales of chip manufacturing and AI model training can be done. India’s difficulties in implementing its chip-manufacturing vision via its PLI scheme is reflective of the underlying challenges.

Thus, AI training and deployment face bottlenecks that are hard to resolve, and this also bodes well for those worried about their jobs.

Third, not all roles are the same. Even where automation is taking place, the roles that are getting “replaced” are in sectors like finance, accounting, and HR, in roles that are not customer facing. And even in the sectors where roles are being automated, it is not like whole departments are being liquidated.

Those who display abilities to leverage the technology of generative models and other tools to improve their productivity, far from being fired, are getting rewarded through higher positions and better compensations.

Fourth, the impact on jobs is not going to be geographically equitably distributed. The Global North, with its better digital infrastructure and higher skilled employees, is going to see a greater impact due to AI compared to the Global South. And automation is not just going to impact the local job markets of the Global North, but also impact the telemigrants – people living in one country while working for a company based in another.

Thus, on the whole, the jobs debate is a mixed bag and will continue to be so. Some roles will go, some new ones will emerge, and the demands for rising productivity will prevail. The only thing that can be unequivocally said is that mass reskilling and upskilling is going to be needed in this coming age of AI.

IBM’s recent study found a staggering 40 percent of the workforce needed reskilling to benefit from AI. And firms like Wipro are investing over a billion dollars to train their over 2.5 lakh staff in these new technologies. Thus, for now, all that one can say is that AI won’t replace people, but people who use AI will replace those who don’t.