Amy Treichi: AI adoption demands accountability, not just speed and innovation

Artificial intelligence has quietly worked its way into the daily routines of companies and individuals, and that is exactly why adopting it is a question of accountability as much as efficiency. Amy Treichi, founder of Great Lakes Engineering Group, argues that businesses can no longer treat AI as a fringe experiment: it already shapes decisions involving safety, money, and trust.

AI is already inside the workflow

Treichi doesn't describe AI adoption as a futuristic scenario. She describes it as the present. She says you can see it even in small everyday things: streaming recommendations, voice assistants, and ads that adapt to people's conversations and interests. For her, that was the signal that the technology stopped being a standalone tool for specialists long ago. It has become a layer that accompanies almost every digital action.

That leads to her main takeaway for leaders: teams will use AI anyway, even if hardly anyone at the company talks about it. Treichi points to a trend showing that roughly three quarters of companies already use AI, and she reads it as the end of the wait-and-see era. If employees already rely on these tools at work, management's job isn't to ban everything. It's to understand where the technology speeds things up and where it introduces a new class of errors and risks.

Rules, not bans

To understand the subject better, Treichi went back to school herself and completed a five-week intensive program on prompting. She says the experience only sharpened her sense of the scale involved. She compares AI to the arrival of the World Wide Web, but stresses that the current wave is moving even faster. That, she argues, is why companies need clear usage rules rather than digital transformation manifestos.

In its engineering work, Great Lakes Engineering Group already uses AI where it speeds up routine tasks without sacrificing control. The point isn't to hand the final decision to a machine, but to give people a boost at intermediate stages of the work.

• Translating complex engineering briefs into plainer language for clients
• Producing structured meeting minutes in minutes instead of hours
• Collecting and organizing large volumes of project data
• Drafting updates, emails, and internal documentation

Her core principle is strict, though: no AI output should move forward without human review, especially on infrastructure projects, bridges, and transportation facilities. In that environment, model hallucinations aren't an abstract problem. They can lead to a bad decision.

"It helps me as an assistant, and sometimes as an advisor.

But in the end, everything comes back to me. I check it before it goes any further. You can't take your hands off the wheel."

Where the line is

Treichi draws a clear line between useful automation and abuse. If AI clears away administrative busywork, speeds up document prep, or sorts data into order, that's a legitimate productivity boost. But if someone passes off generated work as entirely their own, or bills a client five hours for a task that took five minutes with AI, the problem is no longer the technology. It's professional ethics.

In engineering, and on projects involving public safety or public money, the issue gets even sharper. Treichi believes one team's internal discipline isn't enough here: companies need policies that spell out acceptable and unacceptable use cases. She stresses that leadership in the AI era doesn't mean ignoring new tools. It means quickly setting the boundaries for how they're used.

She also addresses the broader consequences. Because millions of people have access to AI, regulation can no longer be left solely to technical specialists. Treichi believes that as these systems gain influence, lawmakers will need to get involved, because the stakes go beyond business productivity to basic rules for society. She sees the technology's dual nature even in her personal life: AI can help people with communication difficulties, but adults still need to explain that it's a tool, not a person.

What it means

Treichi's main point is simple: the next phase of AI adoption won't be defined by the number of pilots or the speed of automation, but by the quality of oversight. The winners won't be the companies that plug AI into every process first. They'll be the ones that write the rules in advance, keep humans in the review loop, and don't let efficiency become a substitute for accountability.