How to build a strong AI culture

Forget the kids…the employees are not alright, and a poor AI culture may be to blame.

According to a September report from IT software company GoTo, 83% of knowledge workers and IT leaders are scared of being blamed, or even fired, for an AI-related mistake, with 31% saying they hesitate to use AI for that reason.

The fear doesn’t stop there: about one-third (31%) of employees feel an “unspoken pressure” at work to keep quiet about AI-related mistakes. Another 14% said they have reported AI errors to a leader, but were told to keep it under wraps.

What’s all the fuss about? Peter Mahoney, chief commercial officer at GoTo, told IT Brew the reluctance to hold AI accountable for its mistakes is likely driven by a desire to reap productivity gains from the technology—a frequent talking point among executives.

“We’re not seeing a lot of companies say, ‘It’s a bad idea to out AI for making mistakes.’ There’s a lot of pressure from employees to help support the idea that AI is delivering business results because companies want to be seen as being forward-looking, so that’s where [the pressure is] coming from in general,” Mahoney said.

But there are repercussions when company culture drives employees to stay quiet about AI mistakes, he added.

“If it’s a business that has a more significant impact, maybe dealing with financial data or dealing with healthcare, as an example, there could be a really big impact for not identifying if there was an error,” he said.

Building a healthy AI culture. Having a strong AI culture where employees feel safe to report issues is important because of the potential risks of the technology being misused, Mahoney said.

“Having that culture and having the willingness to reach out and tell people if something is not working is…really critical in a business,” Mahoney said.

Mahoney said part of building a good AI culture involves embracing transparency about the technology’s use within the organization’s workflows—a move that can also help pacify anxieties around displacement.

“Just having a clear conversation and an understanding about these tools and where they’re being used is pretty important,” Mahoney said, adding companies should also be receptive to questions from staffers about the technology.

Rishi Bhargava, co-founder of authentication and identity management platform Descope, said that companies should give employees access to the right tools and highlight good AI use cases to encourage adoption as opposed to mandating it. One example could be showing employees how a leader uses Claude to achieve a specific task during a team meeting.

“You are showcasing the use case and showcasing the person who came up with that use case, and everybody feels encouraged to then go use that,” Bhargava said.

It starts from the top down. Just like a company’s overall culture, building a healthy AI culture starts with top leadership, according to Mahoney.

“It’s something that has to come from senior leadership, but then it has to be enforced more broadly and across an organization,” Mahoney said.

He added that a good culture needs to be complemented with written AI policies that employees can refer to whenever they have questions: “Writing it down is a really helpful process for companies to test and understand where they think the boundaries and usage rules should be for AI.”.