Which tech jobs are at lowest risk of automation? That’s a tough question: as artificial intelligence (A.I.) and machine learning platforms become increasingly sophisticated, it seems like more and more jobs could end up the responsibility of automated software—even higher-level programming.

However, Burning Glass, which analyzes millions of job postings via its NOVA platform, has taken a stab at estimating which tech jobs aren’t as likely to disappear in the years ahead. As you can see from the chart below, the majority of these jobs will also enjoy a stunning growth rate over the next several years, which means those technologists just starting their careers, and trying to decide on an occupation, can consider these “safer” than others: 

It’s easy to guess why software developers top this list of tech jobs at low risk of automation: although it seems likely that software will automate some lower-level programming tasks (indeed, the rise of “no code” programming platforms suggests this transition is well underway), elegant and effective coding ultimately requires human levels of logic and creativity. (That computer programmers will bleed jobs in the years ahead—the only job on this list to do so—just shows that developers with a varied skillset will take precedence when it comes to building software.) 

In a similar vein, it’s logical that analysts would have a sizable “moat” against automation. After all, it’s easy for a software platform to crunch even the largest datasets in all sorts of ways; but it takes human intuition—and again, creativity—to actually draw useful insights from all that data. On top of that, analysts must convey a “good story” about those insights to other stakeholders, including a company’s top executives—something a machine simply can’t do on its own. Those kinds of tech jobs don't go away anytime soon.

An interesting occupation on this tech jobs list is “computer user support specialist,” since the conventional wisdom dictates that automation will absolutely ravage customer support over the next decade or so. Indeed, a new report from Wells Fargo suggests that chatbots and other automated software could destroy as many as 200,000 jobs in banking and finance over the next decade, many of them in customer service; other reports have stated that other industries are similarly vulnerable. 

While you can train a chatbot to follow a simple script, though, there are a few issues here. Top among them: human beings simply don’t like interacting with chatbots, even if they’re not entirely sure that they’re interacting with a piece of software as opposed to a human being. That “revulsion factor” may slow adoption, despite the rise of chatbot-building tools such as Microsoft’s Bot Builder SDK (part of its Cognitive Services) and Google’s Dialogflow (formerly API.AI).

Second, there are many types of customer service that can’t rely on a simple script. For example, people tasked with helping customers navigate through complex systems; while automation can assist in that process in some ways (such as early-stage diagnosis), it seems like humans are needed for the more nuanced end-stages, where diagnosing an actual problem may depend more on intuition and pattern-recognition than anything else. Again, these kinds of tech jobs seem to be safe well into the future.

While nobody should underplay the impact of automation in the years ahead, in other words, it’s clear that many tech professions will endure just fine. If you’re just starting out in the technology field, and exploring your first tech jobs, take heart that skills such as good communication will continue to be valued, no matter how sophisticated the machines around us become.