Companies' Responsibilities in the Age of AI

Spend a minute on LinkedIn and you'll find an executive sharing an internal memo about AI expectations. They mostly follow the same script: AI is now mandatory across the tech stack, it's not meant to replace employees, headcount growth needs to be justified against what AI can already do, and the company is trimming its contractor roster. Read closely and the tension between those last three points is obvious. These memos have two audiences: employees expected to treat them as gospel, and investors reading between the lines for a future-proof strategy.

AI will shrink headcount over time. Companies will prefer to get there through attrition rather than layoffs, because attrition doesn't show up in a press release. That doesn't mean the shift is being handled well. Five things I'd want any company serious about this transition to do:

Invest in training, on the clock. Some companies treat AI adoption as the employee's problem to solve on their own time, on the theory that it separates the motivated from everyone else. There's some truth in that, but it's still the company's job to give every employee a real shot at learning it as part of the work, not a side project.

Don't create an AI imbalance. If a company uses AI to parse resumes and screen candidates, cutting its own workload, it should be doing the same for applicants, not running a recruiting process with a creaky form that makes people retype the same information three times. And if that same company is telling candidates not to use AI on their cover letters, someone should ask why the rule only runs one direction.

Reconsider the no-consultants stance. I'll disclose the obvious bias here. But the logic still holds: a co-pilot surfaces information about things that have already been done, because it's trained on what already exists. A consultant worth the rate comes up with something that hasn't been tried yet. AI replacing the best independent advice isn't a given; it's an assumption a lot of executives are making because it's convenient.

Stay in the conversation about where this goes. The most confident AI-as-the-future-of-work voices think a company can run on a handful of senior people, each directing a fleet of models where a junior team used to be. If that's right, the effects on the entry-level job market are already visible, not hypothetical. Data entry, cold calling, market research, meeting summaries: the exact activities that used to teach a 22-year-old how a business works are the first to get reassigned. Losing that pipeline has consequences beyond any one company's quarter.

Don't mistake cost savings for the whole return. Reducing headcount and boosting margins is a real mandate under capitalism, and AI is a legitimate lever for it. But a company that eliminates its own workforce loses the people who used to vouch for it, in the market and to future hires. A company with no employees left to speak well of it, and no way to explain a stall if the technology plateaus, has traded a durable asset for a short-term one.

I use AI daily. I work with AI companies. I don't think any of this is a reason to slow down. But a lot of companies are moving fast on the returns without thinking hard about the second-order effects, on their own workforce and on the broader labor market they're still operating inside of. That gap is where the real risk sits, not in the technology itself.