AI’s Workforce Impact in Numbers: What FE Leaders Need to Know

For years, speculation has dominated discussions about AI’s impact on jobs—will it replace workers, or simply enhance productivity? Now, new data from Anthropic provides the clearest picture yet of how AI is truly being used in the workforce.

A landmark study by Anthropic, published on 10th February 2025, moves beyond forecasts to reveal how AI is already embedded in professional workflows. Drawing from over four million anonymised interactions with Claude, the study offers the clearest picture yet of where AI is accelerating efficiency, reshaping professional roles, and—crucially—where its impact remains marginal.

Claude is an AI assistant developed by Anthropic, similar to OpenAI’s ChatGPT but designed with a focus on safety, steerability, and long-context understanding. While ChatGPT is widely known for its general-purpose AI capabilities, Claude is often used for deep analysis, coding, strategic reasoning, and long-form writing, making it a valuable tool in professional settings.

This study is the focus of MKAI’s (Morality and Knowledge in Artificial Intelligence) Executive Briefing on AI Usage in the Workforce, which distils the findings into actionable insights for FE leaders, helping translate AI’s workforce impact into practical changes in teaching and qualifications Read the MKAI report here.

For FE leaders, the shift from speculation to empirical evidence brings a turning point: How can training models reflect how AI is actually being used in the workforce today? With concrete data now available, discussions about AI’s impact on jobs no longer need to rely on assumptions. This evidence offers a chance to shape qualifications, teaching approaches, and career pathways with a clearer understanding of AI’s role—focussing educational outcomes on the reality of work today, not just predictions about its future.

Key Insights from Anthropic’s Study

Anthropic’s findings reveal:

1. AI is an Augmenter More Than an Automator

• 57% of AI usage supports human workers rather than replacing them.
• 43% automates tasks with minimal human input.
• AI is emerging as a collaborator, enhancing professional judgment, efficiency, and problem-solving rather than displacing roles outright.

Implications for FE:

Graduates must be prepared not only to use AI but to work alongside it—refining, validating, and improving AI-generated outputs as part of professional workflows.

2. AI is Reshaping Mid-Level Professions the Most

• The highest adoption rates appear in mid-to-high-wage professions, particularly in finance, consulting, and software engineering.
• AI usage declines at both extremes—less prevalent in low-wage jobs (e.g., retail, hospitality, construction) and among high-wage specialists (e.g., doctors, senior executives).

Implications for FE:

The most profound changes are occurring in middle-tier professional roles—the very jobs that many FE students will enter. Training that reflects AI-augmented workflows rather than simplistic ‘automation or replacement’ narratives is now essential.

3. Where AI is Making the Biggest Impact—and Where It’s Not

• Software development (37.2%) and writing/editing tasks (10.3%) see the highest AI adoption rates.
• Industries reliant on physical labour—healthcare, construction, transport—remain largely unchanged.
• Just 0.1% of AI interactions analysed related to agriculture, farming, or forestry.

Implications for FE:

AI fluency must be embedded in digital, business, and technical disciplines, where AI is already transforming workflows. Equally, manual and people-centric professions remain deeply human-led, underscoring the continued importance of skills that AI cannot replicate—complex decision-making, ethics, and interpersonal communication.

AI is Reshaping How Work is Done—Not Just Which Jobs Exist

The most significant shift is not in which jobs AI replaces but in how work itself is changing. Professionals are offloading analytical and repetitive tasks to AI, using it to refine thinking, accelerate decision-making, and improve creative outputs.

For education, this raises critical questions:

• Who gains the most from AI-driven productivity?
• Which new competencies will be essential in an AI-integrated workplace?
• How can qualifications keep pace with these shifts?

FE colleges are positioned to play a strategic role—not by reacting to AI’s impact after the fact, but by leading the charge in equipping students with both the technical proficiency and the human adaptability required to thrive.

Strategic Priorities for FE: Preparing for an AI-Augmented Workforce

1. AI in Digital and Business Education

AI is already indispensable in business, marketing, and IT. Graduates entering these sectors must develop a sophisticated understanding of AI’s role—how to refine, challenge, and oversee AI-generated work rather than simply use AI tools.

2. AI in Technical and Vocational Education

Professions such as engineering, finance, and professional services are undergoing rapid AI augmentation. Curricula should move beyond introductory AI concepts, integrating data analysis, automation strategies, and AI-assisted problem-solving to reflect how work is changing.

3. AI in People-Centric Professions

Fields like health, social care, teaching, and customer service remain fundamentally human-led. The focus here is not AI substitution but working alongside AI-driven decision-making systems—whether in clinical diagnostics, HR analytics, or adaptive learning technologies. These professions require a renewed emphasis on ethical reasoning, leadership, and human-centred problem-solving.

What FE Colleges Can Do Today

Rather than reacting to AI’s workforce impact after the fact, FE colleges have an opportunity to lead. With AI already transforming professional roles, the focus must shift to preparing students for AI-augmented work environments. Practical steps include:

Audit existing curricula for AI integration

Review current courses to assess where AI is already relevant and where its impact is growing. Ensure qualifications reflect AI’s role in digital, business, and technical fields.

Engage with employers to track AI-related workforce changes

Partner with industry leaders to understand how AI is reshaping job roles and workflows. This will help align training with real-world demands and emerging career pathways.

Invest in staff CPD focused on AI fluency across disciplines

AI isn’t just for IT courses—it is influencing business, engineering, healthcare, and more. Teachers need professional development that equips them to embed AI knowledge into their subject areas and guide students in using AI effectively.

By taking these steps now, FE colleges can ensure their graduates are not just AI users but professionals who can work effectively alongside AI, shaping its role in their industries rather than being shaped by it.

How Reliable is the Data? Examining Its Validity and Limitations

The Anthropic study provides one of the most detailed empirical analyses of AI’s role in work to date. However, like any dataset, it is not without its limitations, and these must be considered when interpreting the findings.

One key concern is “Measuring Inside the Bubble.” The study is based entirely on interactions with Claude, Anthropic’s AI model, which means it does not account for AI usage in other widely used systems such as ChatGPT, Gemini, or Mistral. This creates a potential blind spot—are the observed usage patterns representative of AI adoption across the broader workforce, or are they unique to Claude’s user base?

Additionally, while the data was anonymised, we lack critical demographic details about the users:

• Who are they? Are they concentrated in specific industries, geographies, or income levels?
• What is their intent? Are they using AI for daily professional work, occasional assistance, or simply experimenting?
• How representative is this sample? Without understanding the user base, it is difficult to determine whether the findings generalise to the workforce at large.

Like any dataset, there are gaps in what it reveals. We don’t fully know who these AI users are, how representative they are of the broader workforce, or whether these adoption patterns will hold over time.

Another dimension is the broader challenge of preparing individuals for AI-driven work. Should education focus on teaching specific AI tools, or on the deeper skills needed to work alongside AI? The risk of shaping qualifications around findings from a single dataset is that AI adoption patterns may change, making today’s insights less relevant tomorrow.

While the Anthropic study is valuable in offering real-world data, it is not a definitive map of AI’s impact across the workforce—it is a snapshot of one segment of usage at a particular point in time. For education leaders, this means using such findings as a guiding input, rather than a rigid blueprint, and continuing to assess AI’s role in work through ongoing industry engagement and real-world observation.

FE Must Lead, Not Follow, in Workforce Adaptation

Reflecting on this study, FE leaders can now see that AI is no longer an emerging trend—it is already a deeply integrated force in professional work. This presents a pivotal opportunity: to prepare students not just with AI knowledge but with the adaptability to work effectively in an AI-augmented world.

Rather than focusing solely on ‘AI skills’ or treating AI as just another tool, the conversation needs to shift towards a broader understanding of how AI is reshaping the way work is done. Many professionals now have an AI-first mindset, seamlessly integrating AI into their workflows to brainstorm creative ideas, refine their thinking, challenge assumptions, and accelerate their output. While some employees rely on company-approved AI, many use shadow AI—unofficial tools outside organisational oversight—to increase efficiency.

The specific tools matter less than the way they are used. Workers are moving fluidly between Claude for coding, ChatGPT-4o for content creation, Perplexity for research, and Google Gemini or Microsoft Copilot for productivity within enterprise systems. AI is no longer a separate function—it is becoming second nature to many in the workforce.

To put this into perspective, I recently asked ChatGPT to estimate how many questions I had asked since November 2022, and it suggested at least 50,000. A single business strategy conversation I copied into Word spanned 114 A4 pages and 35,000 words. This is the scale at which AI is being embedded into professional thinking and decision-making.

For FE leaders, the challenge is clear: graduates must not be playing catch-up. The gap is widening between those who can harness AI effectively and those who cannot. Understanding how AI is already being used in the workforce—and embedding this reality into education—is no longer optional.

About the Study

This article is based on Which Economic Tasks are Performed with AI? Evidence from Millions of Claude Conversations, a study published by Anthropic on 10th February 2025.

Authored by Kunal Handa, Alex Tamkin, Miles McCain, Saffron Huang, Esin Durmus, and a wider team of researchers, the report analyses over four million anonymised interactions with Claude AI.

It provides one of the most detailed, data-driven assessments of AI’s role in professional work today, offering valuable insights into how AI’s integration into the economy is evolving.

Access the report from Anthropic here: Anthropic Economic Tasks AI Paper

By Richard Foster-Fletcher, Executive Chair of MKAI