AI won’t steal your job, but you need to learn to work with it

The truth is, we’re still very early on in the AI revolution, if that is to be what this becomes. We’re not entirely sure how or if each industry is going to find those efficiency gains, what issues are going to crop up as a result, and what work will look like on the other side. But each tech giant and a million smaller companies are desperate to make sure that they own a big part of it.

Rise of the AI agents

Jamil Valliani, Atlassian’s head of product AI, watched AI grow throughout his 20-year stint working on Bing at Microsoft. He said a large amount of current workplace AI products are just feeling things out and have low overall usage, but that Atlassian and others were winning over millions of users with more practical AI, including virtual personnel.

“I don’t think anyone is hired to go write reports, like weekly reports, but it’s a fact of life that you have to do it. Nobody’s hired to go triage all the bugs and organise them into a table, their job is actually to go fix the bugs,” he said.

“When you look back and think ‘what have I been hired to do here?’, it’s often not those tasks. Our hope is that AI team members accelerate those things.”

Atlassian’s AI-focused Rovo software is included with subscriptions to workplace software like Jira and Confluence, pulling in all the data from a company’s “Teamwork Graph” to power personalised solutions through search and chat interfaces. But it also provides Rovo Studio, where teams can create information hubs, automations and even AI workers. These agents can be given capabilities and knowledge, and assigned roles to undertake based on the workplace’s data. The agent could serve any role, from comms person to report writer to IT support to bug triaging.

Two concerns immediately come to mind when considering this kind of integration.

First is the possibility of certain workers being entirely replaced by AI, if for example the entire “thing they were employed to do” really was reports and analysis.

The second concern is that if the AI has been trained on a massive pool of human-created information, and then we start relying on the AI to produce more and more of our work, aren’t we poisoning the well? Won’t future AI just be training from AI?

Valliani doesn’t see it that way. He said the creativity and ingenuity and decision-making would continue to come from humans, they just wouldn’t need to do as much admin alongside it.

“As we get more and more time to think, and not just do rituals or do processes that are more about managing work, I think we’ll actually see an elevation in thinking, not a reduction in it,” he said.

“As we went from typewriters to word processors, as we went from PC to mobile, from low- to high-speed internet, those things ultimately saved us time and made the wait of ‘dealing with the machine’ a lot less. It gave us more time to do more things and be more powerful. That’s what I expect here. The machine taking care of the more mundane but important background tasks.”

But that’s speaking about the future. Right now, I’d wager the majority of knowledge workers and office workers wouldn’t be comfortable working with Rovo, meaning the context they’re likely to experience AI agents in is as an imposition from somewhere else in the business.

Valliani said the goal would be to integrate AI tools into each team member’s everyday flow, and that the likes of Atlassian was primed to do that thanks to the graph that gives its software so much insight about how people work. But there’s surely a diversity of opinions among workers on whether and to what degree they’d like to have their flow improved in this way.

Get news and reviews on technology, gadgets and gaming in our Technology newsletter every Friday. Sign up here.