Amazon’s AI push runs into employee friction

Amazon is pushing employees to adopt artificial intelligence in their daily work, and according to a report from The Guardian, some workers say their performance is now being judged on how well they do it. At the same time, employees said that they worry that the tools they’re being asked to use could one day replace them.

In the report, Amazon denied mandating the use of AI tools for its workers. However, employees anonymously interviewed said that the pressure to use automated tools that were allegedly developed in a rush is creating more work as opposed to lessening it.

“In the end, my guess is that the developer cycle is not going to change, and [could] even be potentially longer,” said Denny (a pseudonym for an Amazon employee). “This pressure to use [AI] has resulted in worse quality code, but also just more work for everyone.”

“I would get shown these random tools by my manager who’d be like: ‘Why don’t you try using this thing?’, and it was just the result of a hackathon,” said Denny. He says the tools are “half-baked” and unhelpful, and in fact add to his workload because he has to vet them.

The disconnect described by employees reflects a common challenge in large-scale tech rollouts. Chris Gee, founder of Chris Gee Consulting, told Ragan that pressuring employees to use AI tools without clearly defining how success looks, or how they’ll be supported can quickly create confusion and resistance rather than meaningful uptake.

“Before you can realistically roll out a mandate, communicators have to have supports in place to explain the rationale,” Gee said. “Otherwise it’s not going to work, and there’s a much higher likelihood it backfires and creates pushback.”

An AI mandate needs a comms roadmap

While Amazon’s approach to AI adoption has sparked concern among employees, Gee told Ragan that not all aspects of the company’s AI push are misguided from an internal comms point of view. He said that setting a clear expectation that AI adoption will play a role in performance is a necessary step as organizations navigate rapid technological change.

“I think it’s only realistic to say that an organization has to accelerate AI adoption, and to evaluate people based on how well they adopt and integrate it into their workflow,” he said. “Communicating the importance of that makes sense given where we are with this technological revolution that is going to impact all of us in different ways on the job. But there has to be another side of that — there has to be a realistic understanding of what adoption actually looks like, and how long it takes for people to get there.”

However, AI adoption pushes like Amazon’s can fall short when expectations aren’t communicated clearly and workflows are disrupted without proper outreach from the company to affected employees.

“Part of what is driving the pushback is people are saying that it’s taking them longer, and that’s only to be expected,” Gee said of Amazon’s AI adoption efforts. “When we think back to other technological shifts like the internet, it didn’t immediately speed up productivity. At first, the tools were clunky, not everyone was using them and the use cases weren’t clear. It was only over time, as adoption grew and people understood what worked and what didn’t that productivity really increased. I think we’re seeing the same thing with AI. Right now there’s a lot of experimentation, and during that period you’re going to see some slowdown.”

Moments like this are where messaging shifts from explaining change to actively guiding it.

“So expectations and communication have to account for that, otherwise you create a disconnect between what leadership is saying and what employees are actually experiencing,” he told Ragan.

Comms pros should create a priority list for AI adoption comms to bridge disconnects, including:

• Clearly define successful adoption. Be clear about how AI uptake is being measured and tell employees what success entails in detail.
• Normalize the learning curve. Acknowledge that productivity might temporarily dip before it bounces back during AI adoption processes.
• Communicate the impact to individual roles at the outset of adoption. Create messaging about what AI will and won’t do before employees start filling in the gaps themselves.
• Build out visible support structures. Communicate expectations from leaders alongside training offerings and ongoing reinforcement campaigns.
• Close the perception gap early. If employees experience adoption friction while leadership promises increased efficiency from AI, trust is bound to erode.

Gee said this clarity doesn’t happen without internal comms.

“I think the first step should be a really transparent communication about how leadership is thinking about implementing AI and what we see this looking like,” he said. “It’s okay to say this is going to reshape how we think about roles and productivity, because realistically it should. But you also have to explain what that means over time, and what it means for employees in their day-to-day. Otherwise, that’s where the uncertainty comes in, and that’s what ultimately slows adoption.”

Sean Devlin is an editor at Ragan Communications.