There’s a narrative circulating right now that AI is about to wipe out entire industries overnight. But when you look at the actual data, the story is far more nuanced. One of the most interesting reports released recently comes from Anthropic. Yes, the company behind the clawed AI models. And what they’ve done is something pretty unique. They didn’t just speculate about the future of work. They analyzed millions of real AI interactions to understand which types of work AI is actually being used for today. And the gap between what AI could theoretically do and what people are actually using it for is enormous, which tells us something important about the future of work. This is Shaker Unfiltered, where we break down the full world of recruitment marketing from talent attraction and employer branding to martekch and analytics. Join us for real stories, bold ideas, and practical insights that help you stay ahead and take your talent strategy further. Let’s get into it. >> Hi, I’m John Graham, vice president of innovation, inclusion, and growth at Shaker Recruitment Marketing. I’ve spent more than a decade working in recruitment, marketing, and employer brand, helping organizations navigate major shifts in the labor market. And today, we’re going to break down a few things. First, what Anthropic’s latest research reveals about AI and the labor market. Next, we’ll dig into which sectors are most exposed to AI disruption. Also, we’ll look at where the technology is actually being used today. And most importantly, what talent leaders and executives should be thinking about right now because the biggest risk organizations face in this moment is not automation. It’s misreading the shift. Let’s get into it. The anthropic report is called labor market impacts of large language models. We’ll put the link in the description uh so you can check it out as well. Um and we’ll throw some graphics up as we go along. But the key thing to know here is instead of forecasting hypothetical disruption, Anthropic looked at millions of anonymized interactions within their own AI models. They mapped those interactions against the US Department of Labor’s occupational task database, which categorizes thousands of job tasks across industries. From that analysis, they created two important metrics. The first theoretical AI capability. This measures the share of job tasks that large language models could theoretically perform. Second, uh the metric is observed AI usage. So this measures how often people are actually using AI for those tasks today. So effectively what AI could do versus what people are actually doing with it. And that gap is where the real insight and the story lives. So when we look at the data, a very clear pattern emerges. AI adoption is concentrated heavily in knowledge work. This we know. But let’s look at some of the top sectors where AI capability is highest. Top categories include computers and math at 96% of tasks theoretically addressable. Next is business and finance at 94%. Management 92%, legal 88% and arts and media at 85%. These are the fields where work involves writing, analysis, communication, information synthesis, problem solving. You know, these are the types of cognitive tasks that LLMs or large language models excel at. But here’s the important nuance. When we look at actual AI usage, the numbers are dramatically lower. For example, computer and math roles show today 32% real usage. Business and finance 28%. Management 25%. Legal 15%. So even in the industries most exposed to AI, real adoption is still far behind where theoretical capability lies. And that tells us something extremely important. The limiting factor right now isn’t the tech. It’s workflow integration. It’s trust. It’s organizational chain. So now look at the opposite end of the spectrum for a second. Jobs with the lowest AI exposure include construction, agriculture, installation and repair, grounds maintenance, food service, personal care. These are the jobs that involve physical work, right? realworld environments, human interaction uh that AI can’t get to yet. But the large language models simply don’t operate in these environments today. So the takeaway really is AI is not disrupting the labor market evenly. It’s targeting cognitive work first. And this is one of the most significant shifts we’ve seen in the modern economy because historically automation targeted manual labor, right? You think uh your factories, your uh automotive uh you know manufacturing uh assembly, things like that where they put robots in play. But today AI is targeting knowledge work. So another uh critical insight from the anthropic report is how AI is being used. It showcases that in about 57% of interactions AI is being used for augmentation. Meaning the human is still doing the work but AI is helping them do it faster or better. Most of folks that I know are still using it to enhance their email outputs or coalate a bunch of jumbled thoughts into something meaningful, right? Um and in some cases even the automation uh through agentic uh activities and so forth. But this these examples today right where under augmentation umbrella is um some of those examples include like drafting content like I said summarizing documents that um you know are really long reads uh generating code and we’re seeing a lot of uh programmers today writing less and less code because AI just does it better um or even uh information analysis again large data data sets. Uh AI knocks these things out for breakfast. But in about 43% of cases, AI is being used for automation. So where the system is completing tasks with minimal human involvement, that’s the future. This balance between augmentation and automation is extremely important for a few reasons, but mainly it suggests that in the near term, AI is more likely to reshape jobs than eliminate them. The most fascinating part of the report may be the gap between theoretical capability and observed usage across almost every industry. The blue bars, what AI could do are dramatically larger than the red bars, which is what it’s currently doing. This tells us three things. First, organizations are still in the experimentation phase. Second, most companies haven’t redesigned their workflows yet. And thirdly, the real transformation is still ahead of us. The technology is moving faster than organizational change. And historically, this is exactly how every major technological shift unfolds. So, what does this mean for you as a talent leader uh and your executive teams? There’s three big strategic questions that we have to ask. First is workforce design. Organizations need to start asking where should AI augment employees and where should it automate tasks. The companies that win will redesign jobs around human plus AI collaboration not just add AI tools on top of existing work. As we say at Shaker adding AI to broken processes only amplifies chaos. Second question that needs to be asked is around skills transformation. The most valuable employees in the next decade will be those who know how to work effectively with AI systems. Prompting, evaluation, AI assisted decisionmaking. These will become core professional skills and that means learning and development strategies have to evolve quickly. Third question, employer brand and talent attraction. We’ve got to start talking about how AI adoption is going to influence where people want to work. Top talent increasingly wants to work at companies that are technologically progressive, investing in new tools and enabling employees to work more intelligently. Organizations that ignore AI risk looking operationally outdated. So, let’s let’s close this out. If there’s one takeaway from the anthropic research, it’s this. The future of work is not about AI replacing humans. It’s about AI changing the nature of work itself. And the organizations that thrive will be the ones that move early to redesign how work actually happens. Because the biggest labor market shift ahead isn’t automation. It’s augmentation at scale. And we’re just getting started. That’s the crazy part. We’re at the bottom of the mountain here. If you found this breakdown helpful, go ahead and subscribe for more conversations about AI, talent, strategy, and the future of recruitment marketing. And if you are a talent or business leader trying to understand how these shifts impact your workforce strategy, this is exactly the conversation we should be having right now. So, give us a shout.
Back to Episodes