The Social Impact of Gen AI: What is the downside of generative AI?

Technological progress comes with unintended consequences. At the dawn of the Internet, we could easily foresee its benefits. Sharing information on a global scale would provide limitless resources to better ourselves. It would drive huge economic efficiencies, help advance knowledge, bring together global communities, promote rational discourse, and establish consensus on important topics.

The downsides didn’t seem so evident at the time. Unfortunately, the Internet has also proved to be an equally effective vehicle for propaganda and misinformation. It has potentially sowed division – with social media platforms serving as an echo chamber to polarize politics. And it has helped propagate harmful content.

Generative AI also has the potential to serve nefarious purposes. AI deepfake technology is already smearing reputations[1] and supporting sophisticated phishing attacks[2]. Armies of AI-chatbots could become ever more sophisticated in swaying public opinion (imagine the next Cambridge Analytica scandal on steroids)[3]. AI is already generating harmful content[4]. And future software viruses and malware attacks will become even more dangerous using AI. Any of the negative aspects of existing technology can become more prolific (and harmful) as AI evolves.

It is also important to note that efficiencies generated by the information age have not been shared equitably. The advent of AI could well accelerate inequality. Goldman Sachs estimates that 300 million jobs could be automated as the result of generative AI.

Which jobs will generative AI automate?

In previous economic revolutions, it was easier to see the benefit of automation. During the industrial revolution, much of the work automated was potentially dangerous (think factory work and mining). During the early information age, much of it was clerical, or just plain boring (think bookkeepers, shop assistants or supermarket checkouts).

Perhaps one of the most troubling aspects of generative AI is that it threatens to displace jobs that are very much coveted within society: knowledge workers, whose skills are acquired after years of specialised education. These are the sort of jobs that traditionally served as a vehicle for social mobility. Those who study hard, excel at school, master their chosen vocation, can go on to acquire respected, well-paid professional jobs as doctors, lawyers, architects, financial advisers, or software engineers. As AI evolves, it is not inconceivable to see such occupations being disrupted and many traditional jobs within those professions disappearing.

How will we replace those jobs?

Of course, new jobs will be created as part of the process. Research from MIT economics professor David Autor found that “60% of today’s workers are employed in occupations that didn’t exist in 1940” [7]. But if AI begins shouldering much more of the brain power involved in knowledge work, how will the dislocation be managed?

Will the new jobs created be any better than the old ones? What will the impact be on real wages across the economy? Will enhanced productivity help everyone enjoy more leisure time? Perhaps Keynes’ 15 hour working week prediction will finally come true[5])! Or will we be left with more “bullshit jobs” (a term coined by American anthropologist David Graeber[6]), and be left competing ever more fiercely for the ones that matter?

Some might argue that these concerns are not relevant for a LinkedIN post. They are more the remit of politicians and economists than businesspeople. But in trying to envision the impact that AI will have on society, it is hard to ignore its potential to exacerbate societal inequalities and divisions.