SURESH BABU: How AI affects carmakers’ workforce is up to us

Artificial intelligence (AI) has long been a rich source of storytelling: from the HAL 9000 computer — the brainchild of Stanley Kubrick and Arthur C Clarke in the film 2001: A Space Odyssey — to the replicants in the epochal blockbuster Bladerunner and the genocidal Skynet programme in the Terminator franchise.

Some of these stories portray AI as an ambiguous presence, or more often as malign, menacing or even murderous, pitted against the ingenuity and pluck of humankind. At their best they entertain and enthral, while asking ethical and moral questions about humankind’s relationship with AI and its role in society.

We have now reached and passed the years in which many futuristic movies were set. And as we do so it’s apparent that the real-life future of AI and its roles in society, business and industry might well be more nuanced and complex than in our storytelling.

Yes, AI has the potential to kill us all: Stanford University has identified runaway AI as having a role in one of five extinction scenarios, with economic collapse, climate catastrophe, a third world war and synthetic biology in the wild (redesigning organisms for specific purposes).

Trond Aren Undheim at the institution’s Existential Risk Initiative says a pitfall of looking at possible future extinctions is essentially planning for past events such as pandemics rather than being open-minded about what new, unprecedented risks the future might bring.

Yes, AI could bring about societal collapse and conflict: Geoffrey Hinton, the “godfather of AI”, told 60 Minutes that he is deeply worried about AI bots being used to spread fake news and stoke outrage, extremism and violence. In essence, he says they are capable of making us believe that truth and facts don’t exist and to stoke tensions at an interpersonal, societal, national, international and global scale.

He notes that the use of AI-controlled battle robots makes it easier to start wars without worrying about casualties on your own side. Hinton is also convinced that AI will soon be smarter than us. What that will result in remains to be seen. A major difference he cites between the threat from AI and climate change, for example, is that the latter has one central solution — move from fossil-fuels’ power-generation to clean energy.

Improve HR

With AI it is not so simple. Perhaps part of the solution, he suggests, is to build strong ethics into AI that provides guardrails to its conduct. Whatever the case, the potential risks warrant our attention. Hinton was a cosignatory to the statement by the Centre for AI Safety, a nonprofit organisation: “Mitigating the risk of extinction from AI should be a global priority alongside other societal-scale risks, such as pandemics and nuclear war.”

AI is a double-edged sword: It can vastly improve the efficiency of HR by, for example, scrutinising Outlook and Teams to gauge employees’ morale. There are AI programmes that can promote diversity, equality and inclusion in new hirings and promotions. The downside is that, poorly used, it can do vastly more harm than good. Former US labour secretary Robert Reich told CNN that it has the potential to discriminate against or disqualify candidates that don’t fit an existing demographic within an organisation.

In interviews, it might not recognise nonmother-tongue accents and so mark the candidate as less qualified or unqualified. So, as with most things, garbage in, garbage out. At worst, says Reich, AI could eradicate decades of gains in workers’ rights. Either way it has tremendous social and political implications and warrants business’ attention and investment: companies unwilling to adapt to it will fail to survive.

Yes, it will affect jobs. The World Economic Forum estimates that 85-million jobs will be displaced by AI in the next four years. The Brookings Institute’s department of labour database asserts that 740 out of 759 job categories face a near-term risk of AI automation. 

What does this mean for carmakers? Rose Khattar, director of economic analysis & an inclusive economy at progressive think-tank the Centre for American Progress, suggests a fundamental step towards business and society adapting to AI is for legislators to put workers at the centre of the issue by steering the creation of AI to complement employees, upskilling and reskilling workers to prepare them for the adoption of AI, and meeting the needs of workers displaced by AI.

“While technological disruption has long been associated with a doom-and-gloom narrative of ‘robots taking jobs’ and mass unemployment that has not been realised, the potential impacts of AI on the labour market could be different from the past. The adoption of AI could create new jobs, complement existing work, raise worker productivity, lift wages, boost economic growth, and increase living standards.

Terrified artisans

“On the other hand, AI could displace workers, erode job quality, increase unemployment and worsen inequities. Whether and how AI benefits or harms workers is not a foregone conclusion — it is up to [legislators] to guide the users and adopters of AI in the right direction,” she says. 

It is clear that AI must be recognised as the fundamental force of change of our time. But business must also see it through the lens of previous revolutions, each of which was initially seen as a threat to workers. The industrial revolution terrified artisans, who feared it would take their jobs. Instead, it changed the world of work and created exponentially more jobs.

The same holds true for carmakers: automated production lines were seen as job killers, but while automation did cut base jobs it enabled manufacturers to increase output and so created jobs as production facilities expanded. AI will change the nature of the workforce, but it won’t replace it. It’s also generally recognised that AI is more of a threat to white-collar jobs than blue-collar jobs, and that must be factored into our planning. 

Those examining AI’s impact on their business will know that many of AI’s limitations are writ large: if you have used Dalle or seen images it has created, you will have noticed hands with too many fingers, or torsos with extra arms. ChatGPT, while the fastest-growing consumer software application yet, doesn’t always provide accurate summaries of requested topics, and its output may have errors in comprehension.

AI’s content learning pool is also heavily biased, based on what it uses for reference and can easily provide content based in discrimination if not checked by a human. It is devoid of human understanding of context and needs human oversight. With coding, and tasks that require less human nuance and more maths, AI is an excellent assistant. It is Microsoft’s Paperclip in the current world. It is the next evolution of Googling something.

It is a complementary tool, and as such shouldn’t be used as the only tool to perform a job. Therein is the crux of keeping workers’ rights safe — using AI to do the menial tasks for more senior jobs should make more time available to do the important stuff.  

Babu is plant manager at Ford Motor Company of Southern Africa’s Silverton manufacturing plant.