While I’m bullish on the upside of artificial intelligence, I strongly believe that all business people need to approach AI with open eyes — acting strategically to realize its economic, social, and cultural benefits while thinking critically to mitigate its economic, social, and cultural consequences.
To that end, if you haven’t seen OpenAI’s report (published in conjunction with researchers at the University of Pennsylvania) about the impact of generative AI on the workforce, it’s worth checking it out. I’ll warn you though, it’s not sunshine and rainbows.
AI Will Impact Nearly Every Occupation
In short, the team behind the report concluded that AI (specifically GPTs and GPT-powered applications) will impact most occupations to some extent. Most significantly, there are entire fields of work that have “100% exposure” to automation by AI. Jobs like writing, designing, web development, journalism, tax preparation, financial analysis, clinical data management, and climate change policy analysis(!) all fall into this category. Does this mean that 100% of these jobs go away? Not necessarily, but 100% of those jobs are impacted according to this report.
A range of other occupations are heavily impacted, even if not to that same degree: Writers, authors, and lyricists. PR specialists. Proofreaders. Translators. Survey researchers. Court reporters. Blockchain engineers. Animal Scientists. It’s an eye-opening list that houses more than a few surprises.
According to the study, higher-paying jobs and those that require formal education or credentials are more susceptible than blue-collar work and manual labor. This isn’t surprising given that the research looks specifically at the impact of GPTs, not at AI overall or even robotics. So, for white-collar knowledge workers, this one hits closer to home.
To be precise, if I’m reading the report correctly, the measure of “impact” is that a human aided by AI can do a given work task in half the time without a loss of quality, vs a human alone. You could certainly interpret this as good news. GPTs could deliver a meaningful improvement in productivity — and that’s good, right?
Some argue that we are in the midst of a productivity crisis that could have dire consequences. On the other hand, the emphasis on personal productivity (at its worst, exemplified by hustle culture) can lead to burnout and exacerbate health issues. So, others argue that AI-driven efficiencies may turn work into a hyperproductive hellscape in which all of us must lean into an AI toolset to boost output.
Regardless of where you stand on this, one point is clear: You can no longer compete in the future of work if you aren’t AI-fluent. And that alone is argument enough to learn ways to leverage GPTs in your work today.
Back to the report — and I’m reading between the lines, to some extent — one clear implication is that most jobs (and the ones that have 100% exposure, most of all) will be irreversibly changed. Again, this could be good, as AI takes on more of the routine, rote, repeatable (and frankly, robotic) tasks that make up the drudgery in what we do. I think there’s an often unspoken risk that GPT-driven efficiencies automate away far more than this, though — eating away at the work most of us would consider strategic, creative, nuanced, and rewarding. The work that is core to what we do and — for many people — core to who we are. In that case, “learn to prompt” sounds glib. Besides, does the world need that many prompt engineers?
Yes, AI Will Take Your Job
The other implication is that many workers will be displaced. Yes, I’m saying it: A robot will take your job. OK, maybe not your job specifically, and certainly not every job. But I believe that in the near term, many jobs are at risk and many workers will be displaced.
Frankly, it’s not so much about whether GPT-borne productivity will eliminate this particular job or that particular job. It’s that organizations that put efficiency and profitability above everything else will invest in GPT-enabled technologies to justify and build structures that require fewer human workers to serve in the functions that are easier (if not better) handled through automation.
But let’s assume this is neither good nor bad. It just is.
We’ll adapt, as we (almost) always have. Much like hunters became farmers, and farmers found work in factories, then factory workers found knowledge work desk jobs, knowledge workers will now… Well, I’m not quite sure. And lest we forget, great swaths of the world’s population have struggled to make these shifts, many people have been left behind (whether by choice or circumstance), the digital divide is wide, and wealth has accrued overwhelmingly to the few (even if, arguably, the global standard of living has risen overall).
But OK. We’ll adapt. But can enough people adapt quickly enough to avoid near-term technological unemployment? I’m not so sure.
I know it’s in vogue to argue that AI will create more jobs than it destroys. Heck, I’ve been using that line for years. Over the long term, this may be true. But in the nearer term, I don’t see how there isn’t significant displacement — and even widespread deskilling — in the workforce. Again, by choice but more likely by circumstance, many will get left behind. The digital divide will get wider still.
Unlike prior revolutions in the nature of work, the mainstream adoption of GPTs is far too fast to allow for the kind of education and upskilling necessary, at the scale and on the timeline we’re looking at. Anyway, who exactly is on tap to do all that upskilling? Employers? Educators? Governments? YouTube scam-boys hyping get-rich-quick schemes? More importantly, for most “regular” people — the people who don’t spend their days sweating this stuff — this isn’t even on the radar.
(If you’re reading this, you’re inside the bubble. It might seem as if LLMs are practically the only thing anyone talks about anymore. Just a few weeks ago I keynoted the annual offsite for the customer service division of a global manufacturer. When I asked how many people had heard of ChatGPT, four people out of nearly 400 raised their hands. Zero had actually tried it. Let that sink in for a moment…)
Almost Nobody is Prepared for AI
Already, this — along with the fact that the for-profit large language models are black boxes, and that the companies deploying them aren’t doing nearly enough to mitigate their potential harms — has led technology pundits, ethical AI academics, and even tech reporters to suggest slowing things down. In truth, this is unlikely to happen. The arms race and gold rush have already begun.
If you go with the measure of impact used in the OpenAI labor report (completing the work in half the time), one AI-augmented worker can do the job of two. While the company that empowers this across its entire workforce would be extremely productive and extremely successful, the emphasis here is on extreme. That company would be an edge case in its industry (likely in any industry). I’d expect that a more likely case will lead to a substantial reduction in the human labor force as the practical impacts of workplace AI are felt. And yes, entire professions could indeed be blinked out of existence.
Take just one “100% exposure” job: the graphic designer. This past week, Adobe announced its own proprietary LLM (Firefly) trained on properly licensed rights-free content, and its plan to integrate GPT-enabled functionality throughout its Creative Cloud suite of tools. As a frame of reference, these tools are the de facto professional standard for corporate creators. It is this kind of innovation that might halve the time it takes for a designer to complete her work.
But tech analyst David Mattin points out that the impact may go much deeper than this. Provide non-design employees elsewhere in the organization with access to the same toolset and you equip the entire organization with an “infinite in-house visual creator.” His verdict? “RIP design and illustration as we knew it.” As AI democratizes creativity, it utterly devalues it as a dedicated career path. Even if Mattin’s take is too extreme in its assertion, it’s not beyond the realm of possibility.
Maybe long term, it’ll be fine. We’ll catch up. We’ll adapt. We’ll redefine work. According to one school of thought, we may be heading toward a post-work future. Media theorist Doug Rushkoff, for example, has long argued that the notion of working for a living will be seen as a blip on the human timeline, and the idea that a person had to have a job in order to earn access to food, shelter, education, and healthcare will be seen as a relic of times gone by.
In a 7-page letter, Bill Gates argues that there will always be plenty of work to go around. And that AI’s encroachment into knowledge work might cause more of us to seek work as educators and caregivers. Important work, to be sure, but work that requires specialized training (which takes time), is high-stress, unattractive to many, and generally (unfortunately) under-valued and under-compensated. Still, his vision is hopeful — one where humans increasingly work not with their heads but with their hearts.
Depending on your point of view, this might be heaven. Or it might be hell.
Meanwhile, we have the practical implications of AI’s impact on work — and what our lack of readiness means for workers. Constellation Research analyst (and my former Dachis Group colleague) Dion Hinchcliffe suggests a timeline in which AI affects hiring this year, displacement sets in by 2024, and is deemed a “serious” issue by 2026 — yet broadscale reskilling lags another two years (2028) and novel solutions to the labor problem and the emergence of new (presumably more equitable) economic models could be a decade or more away.
Needless to say, this is a problem and we’ve been caught flat-footed. But this isn’t to say that we can’t get back on track. Just that it will take time. And it’s the gap in time between when we feel the impact (as early as — well — now) and when we have enough widespread supports to welcome enough people into an AI-powered future or work that creates the near-term challenge.
And, if you look at the details in Dion’s timeline, even this assumes the use of “Super AI” to help tackle the unemployment problem. Super AI… Not Artificial General Intelligence (AGI), which even many leading AI researchers argue may be decades away or outright unachievable, but the next level above that: artificial super-intelligence (unless Dion is using the term more loosely).
Wither the Firm?
But wait, there’s more.
Jobs provide one lens for viewing the impact of AI. But in a LinkedIn thread about the OpenAI report Peter Evans, chief strategy officer at an e-commerce consultancy, notes “we should also consider firms, since this is how most jobs are organized. When industries are hit by disruption, some firms adapt and grow, while others perish along with the jobs they supported. When we look at the list of jobs that are 100% exposed and see tax and accounting so high on the list, one wonders which, if any, of the Big Four professional services firms (Deloitte, EY, PwC, and KPMG) will remain in 2030-35? If they do, how many people will they employ and what services will they provide?”
It’s a good point. And through this lens, Dion Hinchcliffe’s novel solutions and new economic models can’t come too soon.
As important as it is to consider the future of the firm, it’s more important to consider the responsibility of the firm — and more specifically, the responsibility of the leaders in the firm. All of this is only an issue if the people in charge of who gets to work and in what roles allow themselves to be seduced by AI-driven efficiency above all. In other words, given the inevitability of the human-AI workforce (it’s already here) it remains incumbent on leaders to ensure a human-forward workplace.
What Can Leaders Do to Prepare for AI?
Don’t panic. Do act.
As the impact of AI on work becomes increasingly apparent, business leaders must take action to prepare themselves, their companies, and their workers for the future. Here are some things you can do, starting right now:
- Be proactive in understanding the potential impact of AI on your industry and business, and identify areas where AI can be leveraged to improve productivity and outcomes. Consider all of this through the lens of your human workforce — What changes? What remains the same? Who feels emboldened? Who might feel threatened? What roles are at risk? Which roles will reap rewards? How can you build and sustain a work environment that is human-friendly and human-first, even as it becomes more digitally enabled by the day?
- Offer targeted reskilling and upskilling programs that address the specific skills and competencies that will be in demand in the new world of work. This could include training in areas such as data analytics, artificial intelligence, human-machine collaboration, and other emerging technologies. You can work with educational institutions and training providers to develop programs that meet the needs of their workforce and align with their business goals.
- Redefine roles and responsibilities to take advantage of the unique strengths of both humans and machines. For example, workers could be trained to work alongside AI systems to improve productivity and efficiency, or to focus on higher-level tasks that require creativity, critical thinking, and problem-solving. This could involve a shift from traditional job roles to more flexible, project-based work arrangements that allow workers to develop new skills and adapt to changing business needs.
- Foster a culture of continuous learning and development, where workers are encouraged to take ownership of their own professional growth and development. Business leaders can provide resources such as online training courses, coaching and mentoring programs, and networking opportunities to help workers stay up-to-date with the latest trends and technologies in their field. This can help build a more resilient and adaptable workforce that is better prepared for the challenges and opportunities of the future.
- Consider new business models that leverage the unique capabilities of AI to create new products, services, and revenue streams — and that also place value on uniquely human competencies.
- Collaborate with government and other stakeholders to address the societal and economic impacts of AI, such as job displacement and income inequality.
While the challenges posed by AI on work are significant, it’s important to remember that this is not the first time that technological innovation has disrupted the workforce. With careful planning and thoughtful leadership, it’s possible to create a future in which humans and machines work together to create a better, more equitable society.
Let’s approach this future with open eyes and a commitment to building a better world for all. And hey, let’s be careful out there.
