Artificial intelligence (AI) is changing the way we work. New tools and systems are making powerful algorithms that allow machines to learn from experience or human input to streamline processes without a team of programmers.
Social-media management, e-mail marketing tools and text-to-image generators are among the functions fuelled by AI, representing new frontiers for companies.
“AI is no longer the special purview of PhDs and specialized programmers,” says Stephanie Holko, director of project development at Next Generation Manufacturing Canada (NGen), a Hamilton-based non-profit that supports technology adoption in manufacturing. “Those folks have developed interfaces and platforms to allow others without these specific skills to access the power of their work.”
‘Programmer’ was among the hottest jobs in any industry. Will it remain lucrative when software can program itself?
For now, AI is still a programming specialty, says Jean-Philippe Roberge, professor in the systems engineering department of the Control and Robotics Lab (CoRo) at the Université du Québec’s École de Technologie Supérieure (ÉTS). A lifelong robotics fanatic, Mr. Roberge’s research is rooted in embedded algorithms and the application of AI in robotics.
But AI’s use in the workplace is becoming more widespread. “We’re seeing more AI tools in integrated development environments (IDE) … in which developers and programmers work, where they write their code,” he explains.
“Advanced IDEs are now starting to be able to complete what programmers are coding,” Mr. Roberge says. “Not only by suggesting how to complete a single line, but by suggesting several lines of code at once.”
These tools aren’t necessarily ushering in the death of the programmer, he adds. General-purpose models and tools such as OpenAI Codex and GitHub Copilot are providing software engineers with greater access to systems that use AI to write code – which saves time and energy. Still, he says, “the programmer would have to do higher-level work, thinking about what the function would be of the code, and how to organize the code, rather than having to write every line of code in their programs.”
Subject matter experts (SMEs) – the people who have been on a shop floor for decades and can tell by listening when something is wrong – can’t immediately be replaced by AI, either, Ms. Holko says. But AI can learn what SMEs know and then apply that knowledge to optimize processes. “In my experience, successful AI applications require the collaboration of [both] the AI experts and the SMEs.”
At NextGen Canada, she’s seen this up close: The organization partners manufacturers of products such as solar cells, nanode batteries and marine-vessel coatings with tech companies that produce tools to make this work simpler.
“In a manufacturing setting, there’s context that’s important – health and safety considerations, engineering parameters, cost implications,” Ms. Holko says. “You need that SME to say, ‘this is the goal and these are the constraints.’ A computer can’t just figure that out. You still need somebody who knows what they’re doing to tell the computer what’s good, what’s bad, what’s optimal.”
One example of AI system collaboration is quality control: A subject-matter expert teaches the low-code system what ‘quality’ means, and defines the necessary constraints such as speed and cost. Although specifics for different industry and machine applications will vary, there are enough similarities to allow no-code AI platforms to be implemented successfully.
“Algorithms are built to efficiently solve a particular type of problem – in manufacturing that’s often process optimization or machine performance,” Ms. Holko says. “Someone who knows a lot about the operation can be a subject-matter expert and decide the input and the output signals and what’s good and bad performance. But they don’t necessarily need to be able to code.”
Advanced AI systems do mean programmer job descriptions are changing. So are those of their teachers. “From what I see as a professor, there’s definitely a shift,” Mr. Roberge says. “Having programming skills is definitely still important. But employers are looking less for an expert in a certain programming language. Rather, they would expect our graduating students to have a certain general knowledge about AI and the main families of AI algorithms.
“We’re shifting to offering more AI courses with applied projects. We try to really reflect what’s going on in the industry right now.”
It could be characterized as a retooling. AI is here, and it’s streamlining the way companies work. As more and more industries lean into automation – and finesse their programming teams accordingly – standards of work will evolve.
“We’re still in the early enough stages that adopting this technology is a competitive advantage, but at some point, the companies that haven’t adopted it will be at a disadvantage,” Ms. Holko says. “The balance will shift.”