Which Jobs Will AI Replace in the Next Two Years?

TL;DR

Routine computer-based work and text-in, text-out roles face the clearest near-term automation risk, including data entry, quality assurance, accounting, legal, design, and editing tasks. Workers can respond by using AI now, continuously acquiring skills, and developing the ability to direct agents, coordinate work, and turn ideas into action, although increasingly autonomous agents may eventually challenge even that advantage.

Transcript

what jobs are going to be made redundant in a world where I am sat here as a CEO with a thousand AI agents right i was thinking of all the names of my of the people in my company who are currently doing those jobs i was thinking about my CFO when you talked about processing business data my graphic designers my video editors etc so what what jobs a... Read More

Key Insights

• Routine computer-based work is presented as the clearest near-term automation target. Quality assurance, data entry, and operator roles are highlighted because they often require predictable clicking, typing, checking, and processing that emerging AI technologies can reproduce.
• Text-in, text-out work is considered especially vulnerable in unregulated industries. Jobs that receive a message and produce a text or image artifact may be at risk, including portions of accounting, legal work, graphic design, and video editing.
• Healthcare automation is harder to predict because regulation and public acceptance can slow adoption. AI may be capable of interpreting medical information such as MRI scans, but technical capability alone does not determine how quickly healthcare jobs or tasks will change.
• AI adoption is divided by educational background. The discussion states that about 50% of Americans with college degrees currently use AI, while usage is significantly lower among those without degrees, creating conditions for an expanding capability and income gap.
• Automation risk is uneven across workers. Figures cited in the conversation place about 80% of working women in at-risk jobs compared with just over 50% of men, while high school diploma roles are described as more exposed than bachelor's degree roles.
• Business process outsourcing is highly exposed because many outsourced roles involve standardized digital processing. The discussion warns that jobs sent to India and the Philippines may disappear, even though this employment model has already helped millions of people move out of poverty.
• Continuous skill acquisition is presented as essential for adapting to AI. Some outsourcing workers are already transitioning into training AI, and the need for increasingly varied data may sustain a substantial industry focused on teaching and improving AI systems.
• Human agency is proposed as a valuable differentiator after AI reduces the advantage of raw cognitive ability. People who initiate projects, coordinate agents, and build businesses may thrive, but increasingly autonomous agents could eventually perform many of those actions themselves.

Questions & Answers

Q: Which jobs are most likely to be automated by AI soon?

Routine jobs performed primarily on computers are described as the most exposed. The discussion specifically identifies quality assurance, data entry, and operator roles built around predictable clicking and typing. It also places portions of accounting, law, graphic design, and video editing at risk because these jobs often transform messages, data, or instructions into standardized text, images, or other digital artifacts.

Q: Why are text-in, text-out jobs vulnerable to AI?

Text-in, text-out jobs follow a structure that AI can increasingly reproduce: a worker receives a message, processes information, and creates an artifact such as text or an image. When the process is routine and operates in an unregulated industry, organizations face fewer barriers to automation. The discussion therefore treats predictable digital production as more exposed than work constrained by regulation or public acceptance.

Q: Will AI replace healthcare workers who interpret medical scans?

AI may affect medical interpretation tasks, including reviewing MRI scans, but the timing is difficult to predict. The conversation emphasizes that healthcare is constrained by regulation, institutional requirements, and people's willingness to accept AI involvement. These factors can delay economic adoption even when the technology is capable of performing part of the task, so technical feasibility does not automatically mean immediate job replacement.

Q: How could AI automation widen inequality?

AI could widen inequality because access and adoption are already uneven. The discussion states that about 50% of Americans with college degrees use AI, while rates are significantly lower among those without degrees. People who experiment with AI can improve their capabilities and move faster, while workers unfamiliar with the technology may fall behind. That divergence could expand existing educational, geographic, and economic divisions.

Q: Why are women described as more exposed to AI automation?

The conversation cites a reported disparity in the kinds of jobs held by women and men. About 80% of working women are said to be in at-risk jobs, compared with just over 50% of men. It also states that jobs requiring only a high school diploma have an 80% automation risk, while those requiring a bachelor's degree have a 20% risk, indicating unequal exposure across occupations.

Q: What could AI do to business process outsourcing jobs?

Business process outsourcing faces substantial risk because many of its roles consist of standardized computer-based processing that AI can perform. The discussion focuses on work sent from Western countries to India and the Philippines and warns that these jobs may disappear. This shift could be economically severe because outsourcing has provided employment that helped millions of people move out of poverty, even as some workers begin training AI systems.

Q: How can workers prepare for AI-driven job disruption?

Workers can prepare by using AI now, continuously acquiring new skills, and moving beyond tasks that are purely routine. The discussion points to emerging work in training AI and emphasizes the value of directing agents, coordinating projects, starting businesses, and turning decisions into action. This strategy is not presented as permanent protection, since increasingly autonomous agents may eventually perform some of those higher-agency activities too.

Q: What does human agency mean in an AI-driven economy?

Human agency means the ability to initiate action, get things done, coordinate AI agents, create businesses, and give effective instructions to digital workforces. One speaker argues that this quality may matter more than IQ as AI broadens access to cognitive capability. Another notes a limitation: agents may identify market opportunities, build software companies, manage schedules, conduct research, and anticipate needs, reducing the agency humans must personally exercise.

Summary & Key Takeaways

• Routine digital work is presented as the first major target for AI automation. Jobs built around predictable clicking, typing, processing, or producing text and images are considered especially exposed. Examples include quality assurance, data entry, accounting, legal work, graphic design, video editing, and some medical analysis, although regulation may slow healthcare adoption.

• AI adoption may deepen existing economic divisions. About half of Americans with college degrees reportedly use AI, while adoption is significantly lower among those without degrees. The discussion also cites disproportionate risks for women, high school graduates, and workers in business process outsourcing, while noting that some displaced workers are transitioning into AI training.

• The speakers disagree about what remains uniquely valuable. One view emphasizes human agency, including starting businesses, coordinating agents, and directing digital workforces. Another warns that autonomous agents could perform those actions too. The conversation ultimately frames AI as a disruption of human cognitive capacity, with benefits such as wider access to intelligence and communication support.

• Key InsightsService work performed through predictable computer interactions is at the highest near-term risk. The discussion identifies quality assurance, data entry, and operator roles as examples because their tasks involve clicking, typing, and following a repeatable sequence that emerging technologies can reproduce.

• Text-in, text-out jobs are especially exposed when regulation does not restrict adoption. Work that begins with a message and ends with a text or image artifact may be automated, placing portions of accounting, law, graphic design, video editing, and other digital professions at risk.

• Healthcare automation may proceed more slowly because regulation and public acceptance constrain how quickly AI enters medical systems. MRI interpretation is discussed as a technically plausible use case, but the conversation avoids predicting an exact timeline because healthcare has more limiting factors than unregulated digital work.

• AI adoption is already divided by education level. The discussion states that about 50% of Americans with college degrees use AI, while usage is significantly lower among those without degrees, raising concern that early users will advance faster and widen existing social and economic gaps.

• Automation risk is described as unevenly distributed across demographic and educational groups. Citing figures mentioned in the conversation, about 80% of working women are in at-risk jobs versus just over 50% of men, while high school diploma roles face greater risk than bachelor's degree roles.

• Business process outsourcing is vulnerable because much of the work consists of standardized digital tasks. The conversation warns that jobs sent from Western countries to India and the Philippines may disappear, threatening an employment pathway that has helped millions of people move out of poverty.

• Continuous skill acquisition is presented as a necessary response to AI-driven change. Some workers are already moving from outsourced processing roles into training AI systems, and the continuing demand for new data may create work around teaching, evaluating, and improving models, at least while those systems require ongoing training.

• Human agency may become more valuable as AI equalizes access to cognitive capability. People who can initiate projects, coordinate agents, start businesses, and issue effective instructions may thrive, although autonomous agents that identify opportunities and execute tasks could eventually reduce the value of this advantage too.