De-Skilling vs. UpSkilling with AI - What it means and where you want to be

The scale of this shift is striking. Nearly 49% of jobs now involve AI in at least a quarter of their tasks – up from 36% in early 2025. The legal industry, historically cautious with new tech, is starting to catch up. One recent survey found AI usage among legal professionals skyrocketed from 19% in 2023 to 79% in 2024. In other words, most U.S. law firms are now experimenting with tools like GPT-based assistants for research, drafting, and more. Clients are on board too: about 70% of clients say they're neutral or actually prefer firms that use AI, since it can mean faster and cheaper service.

This surge in adoption isn't limited to tech-savvy coastal firms. AI use is diffusing across the U.S. far faster than past technologies. Anthropic's data suggests that if current trends hold, AI usage per person could equalize across all states within 2–5 years – a diffusion rate roughly 10 times faster than earlier innovations like computers or the internet. In short, AI is going mainstream everywhere, and the window for being an "early adopter" is closing quickly. Law firms that have been slow to adapt should take note: the competitive edge will go to those who integrate AI effectively into their workflows sooner rather than later.

Deskilling vs. Upskilling: Two Sides of AI's Impact

When it comes to how AI changes jobs, experts often talk about two contrasting effects. Anthropic's research highlights both patterns:

Anthropic's data provides a fascinating lens on this. It turns out AI is often tackling the hardest parts of jobs first, not the easiest. On average, tasks that Claude helps with require 14.4 years of education, compared to 13.2 years for typical tasks across the economy. In plain terms, AI is doing a lot of the heavy intellectual lifting that highly educated workers used to do. By taking on these knowledge-intensive tasks, AI can produce a de-skilling effect in many roles: after AI completes the "hard" parts of the job, what remains for the human are often the more routine or manual tasks.

Many white-collar jobs are seeing a similar pattern, which one analyst called the underappreciated story of this AI wave: AI is taking the "interesting" parts of jobs, and often what's left is lower-skill work. However, the flip side is that some roles experience the opposite – an upskilling. If AI mostly automates the easy stuff, the human's work ends up skewing toward the harder, value-added tasks. A cited example is real estate managers: AI handles routine admin like updating records and checking rent prices, while the managers spend more time on high-level work like negotiations and strategy, actually raising the skill content of their job.

Lawyers: From Researcher to Strategic Advisor

For lawyers, especially those in law firms, AI is emerging as a powerful junior associate (that never sleeps). Tasks like legal research, document review, and first-draft writing – the bread-and-butter of junior attorneys – are being turbocharged by AI. Need a memo on the duty of care in Delaware corporate law? An AI like Claude or ChatGPT can pull relevant cases and generate a decent draft in minutes.

Crucially, this does not mean attorneys become obsolete – but it does mean the attorney's day-to-day work shifts. With AI doing the first pass on research or generating a contract template, lawyers can devote more time to higher-order lawyering: advising clients, crafting case strategy, negotiating deals, and appearing in court. These are areas where human judgment and persuasion matter.

Notably, Anthropic's data shows that social skills like negotiation and persuasion have "minimal presence" in AI interactions. No current AI can stand up in a courtroom to sway a jury, or calmly counsel a nervous client through a crisis. Those distinctly human aspects of legal practice remain largely untouched by automation. So in the ideal upskilling scenario, a lawyer uses AI to offload tedious research and paperwork, and then spends the freed-up hours on tasks that truly require a lawyer's insight, experience, and personal touch.

However, there are also challenges and risks to manage. One concern is a form of professional deskilling: if new lawyers rely on AI for answers, they might miss out on learning foundational skills. Traditionally, writing briefs and doing research helped young attorneys refine their legal reasoning. Now, firms must ensure associates still learn to think like lawyers, not just like prompt engineers.

The best approach is to use AI as a teaching tool – for instance, an associate can ask AI for a draft, but then critically review and edit it, comparing it against their own research. Senior lawyers should supervise this process, because AI is prone to errors or "hallucinations" (confidently making up facts or citations). The attorney remains the final arbiter of quality and ethics.

Paralegals and Legal Assistants: Automating the Routine, Elevating the Role?

If any roles in law are ripe for automation, it's the support roles like paralegals, legal assistants, and document reviewers. These professionals handle many repetitive and process-driven tasks: combing through discovery documents, summarizing deposition transcripts, compiling case law excerpts, filling in forms and templates, scheduling and proofreading.

This sounds ominous for support staff, but the outcome isn't necessarily mass layoffs – it's role transformation. We're already seeing forward-thinking law firms redefining what their paralegals and legal assistants do. Rather than spending hours on menial tasks, these staffers can evolve into AI facilitators within the firm.

For example, a paralegal might become the go-to person for operating an AI document review tool: they set up the software to analyze a batch of contracts, train it with examples of what to flag, and then vet the AI's output to pass the most relevant info to the attorney. This requires a mix of legal knowledge and comfort with technology – arguably a higher skill set than pure data entry. In this way, the job becomes more interesting and value-added, focusing on managing the process and adding human judgment on top of AI's raw results.

From a practical standpoint, legal support professionals should proactively upskill to secure their place in the AI-augmented future. This includes getting fluent with legal tech platforms and developing a keen eye for quality control. Studies have shown AI can identify a higher percentage of relevant documents than humans alone, but it may also sweep in some false positives or miss context that a human would catch. The human role shifts to editor and analyst rather than raw researcher.

Legal Operations: Efficiency Gains and New Expertise

Legal operations professionals – the folks who manage the business and procedural side of legal services – are often the unsung heroes driving innovation in law firms and corporate legal departments. For this group, AI is a dream come true – it directly advances their mandate to make legal services more efficient, data-driven, and cost-effective.

Anthropic's index confirms that operations and management roles are embracing AI: business and financial operations work accounted for about 5.9% of AI usage in Anthropic's data, a significant share for a single category. In a law firm context, that could involve:

One important insight from the Anthropic Economic Index is that AI's usefulness isn't limited to programmers – modern AI is designed to be used with natural language prompts, so you don't need to be a coder to use AI tools. This greatly lowers the barrier for legal ops teams to deploy AI. They can experiment with off-the-shelf AI services by literally typing requests, without needing an IT department to build custom software.

Practical Takeaways for Law Firms and Professionals

AI is here to stay in the legal profession, and the data suggests it's better to adapt than to resist. Below are practical insights and strategic takeaways – informed by the Anthropic Economic Index and industry trends:

Conclusion: Adapting to an AI-Augmented Legal Future

The data from the Anthropic Economic Index paints a clear picture: AI is transforming the legal profession, but not by eliminating lawyers. Instead, it's reshuffling the tasks and skills that legal work involves. Some lower-level work will be handled by machines, and new high-level work will emerge for humans. In economic terms, we'll see some deskilling of old routines and simultaneous upskilling into new, more complex responsibilities.

The net effect for those who embrace the change is likely positive – higher productivity, more interesting work, and the ability to deliver better value to clients. Those who resist or delay, however, risk being left with an outdated skill set or an uncompetitive business model.

As Anthropic's research suggests, this is more of an evolution than a revolution: the job apocalypse is not here, but the job description is certainly evolving. By staying informed and proactive, lawyers, paralegals, and legal ops professionals can ensure they continue to thrive in the age of AI-augmented law.