Artificial intelligence continues to divide opinion when it comes to the future of work. While some experts predict large-scale disruption, others see the technology creating new industries and occupations. Amazon founder Jeff Bezos recently argued that AI will ultimately create more jobs than it eliminates, suggesting the technology could even lead to labour shortages by accelerating innovation and economic activity.
Against this backdrop, new research from IT Asset Management Group (IT-AMG) provides an intriguing perspective. Rather than asking economists or HR professionals, the company asked Google’s Gemini 3.5 Flash model a simple question: Which jobs do you think AI will replace, and why?
The resulting list highlights occupations that rely heavily on routine information processing, standardized communication, and repetitive administrative tasks. While the exercise is more thought experiment than prediction, it provides useful insight into how AI itself assesses automation risk in the labour market. For Canada, the findings raise important questions about workforce transformation, skills development, and future economic competitiveness.
Data entry tops the list
The occupation AI identified as most vulnerable is data entry clerk. According to Gemini, the role largely involves transferring information between systems, databases, and documents, activities that modern AI tools and automation software can increasingly perform without human intervention.
This assessment aligns with broader labour-market trends. Across Canada, organizations are investing heavily in digital transformation, process automation, and AI-enhanced workflow management. Sectors ranging from banking to healthcare are seeking ways to reduce manual data handling while improving accuracy and efficiency.
For workers currently employed in administrative support roles, the challenge is likely to be adaptation rather than immediate displacement. Increasingly, employers are seeking staff who can manage automated systems, validate outputs, and perform higher-order analytical tasks instead of simple information transfer.
Second on the list are telemarketers, followed closely by Tier 1 customer service representatives. Advances in conversational AI, speech synthesis, and large language models have made it possible for organizations to automate many routine customer interactions. AI-powered chatbots and virtual agents can already answer frequently asked questions, handle password resets, provide account information, and route service requests.
Canadian businesses have been actively experimenting with these technologies. Banks, telecommunications providers, retailers, airlines, and government agencies increasingly employ AI-driven customer service platforms as part of broader digital modernization initiatives. Yet full replacement remains uncertain.
Customers often become frustrated when automated systems fail to understand context or resolve unusual situations. As a result, many experts believe routine inquiries will become automated while human agents focus on more complex, emotional, or high-value interactions.
Bookkeepers ranked fourth in the AI-generated list. Routine bookkeeping tasks such as expense categorization, invoice processing, reconciliation, and financial record maintenance are increasingly supported by machine learning systems and cloud-based accounting platforms.
For Canada’s small-business sector, which accounts for a substantial share of national employment, automation offers opportunities to reduce administrative burdens and improve productivity. However, this does not necessarily imply the disappearance of accounting professionals. Instead, many are expected to shift toward advisory services, financial planning, compliance, business analysis, and risk management. The lesson may be that automation changes jobs more often than it eliminates them entirely.
One notable aspect of the IT-AMG analysis is that several knowledge-based professions appear high on the list. Proofreaders rank fifth. Modern language models can rapidly identify spelling mistakes, grammatical errors, and stylistic inconsistencies. As AI writing assistants become increasingly sophisticated, organizations may require fewer dedicated proofreading resources.
Similarly, paralegals involved in document review rank sixth. Legal professionals frequently spend significant amounts of time reviewing contracts, case files, and evidence. AI systems can now analyse thousands of pages of text in a fraction of the time previously required. This trend is already visible within Canada’s legal sector. Law firms and corporate legal departments are experimenting with AI-assisted document review, contract analysis, and discovery processes. Importantly, regulatory oversight, legal judgement, and client advisory functions remain firmly human responsibilities.
The future of content creation
Perhaps the most surprising entry is commodity content writers, ranked seventh. AI argues that routine blog articles, product descriptions, and marketing copy can increasingly be generated automatically, particularly where volume and speed are more important than originality or specialist expertise. This observation reflects one of the most visible impacts of generative AI.
Across Canada’s media, marketing, and communications sectors, AI tools are becoming commonplace for drafting first versions of content, generating ideas, and supporting research. Yet human expertise remains essential for investigative journalism, scientific reporting, thought leadership, and content requiring originality, credibility, and nuanced interpretation. The distinction may increasingly be between content creation and content curation.
The list also includes market research analysts, retail cashiers, and commercial translators. For market research, AI can already collect and analyse customer feedback, monitor competitors, perform sentiment analysis, and generate reports. Retail cashiers face pressure from self-checkout systems and mobile payment technologies, although widespread adoption remains dependent on infrastructure investments.
Meanwhile, advances in machine translation continue to improve multilingual communication. Tools such as Google Translate, DeepL, and AI-powered language models have dramatically increased translation speed and accessibility.
Canada’s bilingual environment presents an interesting test case. While AI translation quality has improved substantially, many organizations continue to rely on professional translators for legal, governmental, regulatory, and culturally sensitive communications where precision remains critical.
The broader message is not necessarily that millions of jobs will disappear. Instead, Canada’s labour market appears to be entering a period of substantial transformation. According to IT-AMG Chief Revenue Officer Richy George, the findings reveal how “the middle layer of corporate data management is shrinking,” with information-transfer roles facing particularly strong automation pressures.
At the same time, occupations requiring creativity, emotional intelligence, ethical judgement, relationship-building, leadership, scientific expertise, and strategic decision-making remain comparatively resilient. This aligns with numerous forecasts from organizations such as the World Economic Forum, which anticipate significant job redesign rather than wholesale replacement. For Canada, success may depend on workforce reskilling. As AI takes over routine administrative tasks, demand is likely to grow for workers who can supervise AI systems, interpret results, manage complex human interactions, and drive innovation.