AI & knowledge work
AI is coming for the lawyers and consultants and financiers
I saw a tweet from an anonymous account the other day, which parallels my thinking about the future of lawyers, given increasingly powerful AI. But this post isn’t about lawyers, it’s a more general observation about knowledge work. The tweet is here. If you click through to the tweet, you’ll see that the anonymous tweeter relates the story of talking with a “senior partner at [a] major Bay Area law firm”: the lawyer thinks law firms are screwed given advances in AI, and he or she is looking for an in house role, i.e., a role at a corporation as a lawyer, instead of at a law firm. Much legal work, this lawyer believes, will move in house and future generations of ChatGPT or a different generative AI tool will be the substrate upon which legal work is done.
These observations are all very interesting, and from my perspective, they’re directionally accurate. But I think we can generalize these observations to much or all of knowledge work.
Knowledge Work and Expertise-Based Services
Industries that are heavily reliant on knowledge work and specialzied expertise seem particularly vulnerable to AI disruption. Law is one example, where much of the work involves research, analysis, and drafting doucments—tasks that AI is increasingly capable of doing.
Other professional services like consulting, accounting, and even medicine face similar risks, as AI gets better at quickly processing large amounts of information and rapidly generating insights or recommendations. The expertise that these fields charge a premium for may be increasingly automated.
Hourly Billing Business Models
Business models in which clients are billed by the hour are at high risk because AI can dramatically reduce the time needed for many tasks. If a law firm charges hundreds of dollars per hour for work that AI can now do in minutes, that’s an existential threat to their model. Sure, the law firm could try to complement its increased productivity with taking on more client work, but there is a fixed set of high end legal services required.
Aggregators and Middlemen
Businesses that primarily aggregate information or act as middlemen could be disintermediated by AI. Think of how travel agents have been largely replaced by online booking.
With powerful AI, end users will get the information they need directly, cutting out middlemen who profited from privileged access to information and expertise.
Junior and Entry-Level Roles
AI will have an outsized impact on entry-level knowledge work and tasks typically done by junior employees, like research, basic analysis, and first-pass document drafting.
Junior employees can’t yet match the judgment and experience of senior leaders. But if AI can do junior-level tasks faster and cheaper (even if imperfectly), there may be less need for entry-level hiring. This could make it even harder to get a foot in the door and move up in competitive knowledge work fields.
Insufficiently Differentiated Offerings
AI raises the bar for the value knowledge work businesses need to provide to stay relevant. Offerings that aren’t sufficiently differentiated or which don’t rely on unique data will be easier to replicate or beat with AI.