AI is ready to take over your tasks. Your job? Maybe not

Article content

A new form of artificial intelligence is now being used to help complete tasks traditionally performed by humans to streamline employees’ time and value, which is leading to concerns about exactly what they do and whether some people should fear for their jobs.

These AI agents are advanced AI bots that are developed to complete tasks and make decisions without human interference, according to Forbes. They are not to be confused with generative AI chatbots such as ChatGPT that are driven by human prompts.

Article content

These agents are already being used in many major businesses to take on routine, time-consuming tasks, according to a recent report by Bloomberg. Microsoft Corp. recently launched its own suite of AI agents to help sales, customer support and accounting employees, while its subsidiary LinkedIn Corp. has its own AI agent to take on tasks associated with recruiting.

“Researchers are exploring the implications of increased autonomy and improved learning and decision-making capabilities, while leading general-purpose AI developers are working on building agents based on large language models (LLMs) that will be able to complete certain tasks autonomously,” Innovation, Science and Economic Development Canada (ISED) said in a statement. “However, these are still an experimental technology and are not being widely deployed by businesses.”

In September, 10.6 per cent of all businesses in Canada were planning to use AI over the next year, according to Statistics Canada, with 20 per cent of businesses with 100 or more employees saying they were more likely to use AI to complete tasks in producing goods, delivering services, data analytics and more.

Article content

Article content

The agency also said a majority of businesses reported no significant change in employment levels after implementing AI. But while 69.2 per cent of businesses that plan to use AI to produce goods or deliver services expect no changes to their employment levels, 9.4 per cent expect a decrease.

These changes are part of our evolution

Abdel Hamou-Lhadj

Abdel Hamou-Lhadj, a professor at the Sprott School of Business at Carleton University in Ottawa, described AI agents as accelerators that provide maintenance. He said they can streamline processes, reduce manual work and improve speed, and they are widely used for automation in sectors such as customer support, health care, finance, e-commerce and transportation.

He said that in order for the implementation of AI in businesses to succeed, it needs to adhere to three main senses: business, legal and ethical. The main questions that business leaders should be asking when wanting to use these systems are whether they’ll enhance operations and products, whether it goes against the rules of personal information protection, copyright or cybersecurity, and whether it provides services that are socially and ethically acceptable.

Article content

“In the past, we have been worried about many of the technologies that came up. Of course, there are changes that will take place, but these changes are part of our evolution,” Hamou-Lhadj said. “We need to make sure that there are guardrails from the legal and ethical standpoint.”

Julie Yujie Chen, an assistant professor at the University of Toronto’s Institute of Communication, Culture, Information and Technology, said industry has a long history of using automation technology to enhance productivity. AI agents can help companies automatically make decisions while saving labour time and improving efficiency, but they can also be susceptible to bias.

“When it comes to benefits, we have to understand that benefits are relational,” she said. “Sometimes, the benefits to the employers could be at the expense of workers or job seekers because there are extensive studies that have revealed that AI agents used for hiring tend to reproduce the existing biases.”

Chen said AI agents designed for job recruiting look for key words and qualities within applications, and use those associations to determine if a candidate is in good standing.

Article content

Hamou-Lhadj said biases within the data that AI agents are trained on can lead to inappropriate or unfair responses in chatbots. Another reason biases may be present is a lack of transparency and accountability in AI systems, making it challenging to ensure they function as intended.

“A lot of these systems are developed in a way that you can actually not open them and see, so they are opaque,” he said. “If in one instance they say or do something that’s good, can you guarantee that they will do that in the next instance? It’s not sure.”

Chen expressed a similar concern about AI agents that conduct video interviews and detect a candidate’s facial expressions and answers.

“Very few people actually know exactly how those systems make decisions,” she said. “I think that’s a fundamental problem because of the lack of transparency. We don’t know what kind of expression that they would classify by the technology as ‘this person has good people skills,’ as an example.”

Chen said younger generations are becoming more aware of the issues surrounding AI, and that it’s important to educate others about using these systems in the workplace.

Article content

While business leaders view AI agents as helpful tools to cut costs, they also acknowledge how they could lead to concerns about them replacing human employees.

But Hamou-Lhadj said AI agents shouldn’t be a cause for concern since they could lead to further job opportunities to manage and develop these systems, such as prompt engineers.

“The intent of AI in general and of these agents is not to replace humans, but to enhance the way by which they do their work and to allow them to focus on things that are much more empowering,” he said.

Chen said AI agents could possibly change managers’ responsibilities since they’ll perform tasks on their behalf. She believes that the notion of AI agents allowing workers to avoid doing routine tasks is a myth because they often automate only part of the responsibilities.

“Workers need to also spend time monitoring, making sure the machine works well,” she said. “It’s open to debate whether that part is more liberative, because empirical studies have shown mixed outcomes. It really depends on the context and also on the workers themselves.”

Article content

ISED said the federal government is currently working on measurements to ensure AI agents and similar systems are responsibly used. This includes the proposed Artificial Intelligence and Data Act tabled as Part 3 of the Digital Charter Implementation Act (Bill C-27) in June 2022 and the Voluntary Code of Conduct on the Responsible Development and Management of Advanced Generative AI Systems. It also said the federal government intends to create a Canadian AI Safety Institute.

Bookmark our website and support our journalism: Don’t miss the business news you need to know — add financialpost.com to your bookmarks and sign up for our newsletters here.

Article content