Will AI Take Your Job?

How LLMs Are Changing Jobs Forever

Louis Bouchard

Feb 11, 2025 • 8 min read

Watch the video!

The impact of LLMs on human work is complex and varied across different job categories.

Right now, we think LLMs are most likely to automate or significantly augment tasks that are routine, repetitive, and text-or image recognition and interpretation-based. In creative, analytical, and specialized professional services, LLMs are still more likely to assist predominantly human-led tasks rather than replace them.

We also see a natural progression as we scale foundation model capabilities and improve LLM pipeline reliability. We progress from 1) “unreliable intern assistants” that can help with some one-off tasks and data processing but need lots of clear instructions and lots of extra time checking final results to 2) useful “co-pilots,” which can assist us as we work and increase our productivity to 3) “agents” that can be set to perform relatively complex tasks for hours or days at a time (but still need instructions and checking) and finally to 4) fully automated workflows. Some of these come earlier depending on the reliability threshold of the task and its suitability to the capability of current LLMs. At the moment, most LLM applications are in the first or second category, but agent-based systems are more and more popular and improve daily. This will be the biggest shift from a tool to a complete entity but also the most complex shift to achieve.

As LLM technology continues to advance, it’s important for workers across all sectors to adapt and develop skills that complement AI capabilities. This includes focusing on human traits such as creativity, critical thinking, emotional intelligence, and complex problem-solving. And by the way, that doesn’t mean AI cannot help with these skills. It certainly already can, as we see in our course. As worker’s jobs begin to change rapidly, it will be essential to invest in reskilling and upskilling programs to help workers adapt to changing job requirements, just like leveraging computers and Excel rather than paper tables.

Humans will have to learn how to use AI, understand its ever-evolving strengths and weaknesses, and think through how to adapt their workflows to make use of AI capabilities. We think AI-driven job losses are likely in some sectors where AI boosts worker productivity in some workflows of products that are still supply- or demand-limited. However, we think it is more likely you will be replaced by somebody else who has taken time to learn how to use AI effectively than you are to be replaced by AI directly, at least in the foreseeable future. The future of work will involve a collaborative relationship between humans and AI, with LLMs handling routine tasks and augmenting human capabilities in more complex areas.

So here are some of our thoughts on how different categories of human work could be impacted, which could help you decide where to focus your LLM development efforts!

Knowledge Work and Information Processing

LLMs are particularly well-suited for tasks involving information processing, analysis, and synthesis. Their impact in this area is likely to be substantial:

LLMs can quickly process vast amounts of information, identify patterns, and generate summaries. They are likely to significantly enhance the efficiency of researchers, analysts, and journalists. And talking from my own experience, I can confirm it does provide incredible help to my research and writing, which we have practical examples later on in the following videos of the course.
Writing tasks such as drafting reports, articles, and marketing copy can be incredibly aided by LLMs when leveraged in the right way. While human creativity, nuance, style and error checking remain crucial, LLMs can assist with ideation, structuring, and generating initial drafts.
LLMs can handle a large portion of customer inquiries, reducing the need for human customer service representatives, especially for routine queries. Just keep in mind that it won’t replace all of them, but it certainly helps with easier and safer queries.

Creative and Artistic Work

The impact of LLMs on creative tasks is likely to be more assistive:

LLMs can help with brainstorming, outlining, and generating initial drafts. However, human writers will still be needed for original storytelling, nuanced expression, and final editing.
Text-to-image models can assist in generating visual concepts, but human artists will remain crucial for original artistic vision and execution. We still need a good designer eye, but AI helps do much better both for artists to scale up production or creativity, and for us, non-talented people, to produce something that is nice to see!
LLMs can now even aid in generating melodies, lyrics or whole songs, but human musicians will still be essential for creating emotionally resonant, culturally significant and live music. Still, if you are looking for a background song for a video or elevator, it can be a nice and cheap option!

Technical and Specialized Work

LLMs can significantly augment technical work but are less likely to replace humans in these areas fully:

For programming and software development, LLMs can assist with code generation, debugging, and documentation. Cursor or even just ChatGPT are incredible examples of this. Still, complex problem-solving, system design, and ensuring code quality will require human expertise.
LLMs can help with legal work, from research and contract analysis to drafting simple legal documents. However, interpreting complex laws, developing legal strategies, and representing clients in court will remain human-centric tasks. You don’t want an AI to hallucinate cases or fake proofs.
LLMs can easily assist with financial duties like data analysis, report generation, and risk assessment.
They can help in claims processing, policy analysis, and customer interactions.
Lastly, while LLMs can assist in analyzing medical data and suggesting potential diagnoses, final decision-making and patient care will still rely heavily on human medical professionals. For example, in my PhD, I was using AI to help diagnose multiple sclerosis by training an algorithm to detect early lesions in brain MRIs, which we then sent to expert radiologists to confirm the diagnosis. Nothing was fully automated, but it helped save a lot of time for radiologists to flag relevant information.

Physical and Manual Labor

A lot of progress is currently being made on humanoid robots, and in many cases, are now integrating LLMs and transformer models into the architecture. These are still some way away from commercialization, though, and manufacturing capacity will also be a bottleneck to adoption when they are ready. For now, LLMs have limited direct impact on physical tasks, but they can indirectly affect these jobs;

LLMs can assist in process optimization, inventory management, and robotic control systems. However, many physical tasks will still require human workers because of inconsistencies or expensive sensors and hardware systems.
While LLMs can optimize routes and schedules, as in Google Maps, the physical act of driving or operating vehicles will require either human operators or predominantly non-LM-based AI technology for self-driving cars. The main blockers here is regulation and ethical concerns, but not really the technology though it still isn’t fully perfect.

Management and Leadership

LLMs are less likely to replace human managers but can assist in various ways:

They can analyze data and provide insights to aid decision-making processes.
They can help in task allocation, scheduling, and progress tracking.
They can assist in analyzing employee performance data and suggesting improvements.
They can automate and help streamline many of these tasks.

Moreover, a great manager can now do much more than they used to be able to by leveraging LLMs as another type of employee they are managing, giving clear directions and goals; one can leverage a powerful model like GPT O1 to do some quite complex task almost on its own.

Education and Training

LLMs have significant potential to augment educational processes:

They can provide tailored explanations and practice exercises for students.
They can assist educators in developing course materials and assessments.
They can provide 24/7 personalized tutoring support, remembering past conversations and learning styles. Although human teachers will remain crucial for deeper understanding and socio-emotional development, teachers can facilitate their lives by leveraging those powerful tools.

Routine and Repetitive Tasks

LLMs are particularly effective at automating routine and repetitive language tasks:

Like those involving routine data entry.
or roles such as credit authorizers, checkers, and clerks that may see a significant portion of their tasks automated through the use of intelligent systems where hardware and price is often the limiting factor.

Emerging Roles and Opportunities

The rise of LLMs is also creating new job categories and opportunities:

There will be a growing need for professionals to develop, fine-tune, and maintain LLMs, whether they be AI engineers, LLM developers, ML engineers, or others.
Specialists who can effectively design prompts and end systems to optimize LLM performance.
Experts ensure that LLMs produce ethical and unbiased content.

While it may seem like many of our jobs can be augmented or automated, it doesn’t mean AI will replace us. It simply means we can do better. A recent study compared ****a firm not adopting any AI versus another one adopting and fully using AI for commercial product orientation. They found that the first one will have a 6% probability of job loss in the next 3 years, while the AI-first one will have an employment-positive prospect increased to 16%!

Preparing for the impact of AI adoption

The rise of generative AI is transforming industries, nations, and individuals, presenting both vast opportunities and significant challenges. The key question is: how do we harness AI’s potential while addressing its risks?

Maximizing AI’s benefits requires investments across the value chain — developing energy-efficient chips, advancing cloud infrastructure, and funding innovative AI applications. Governments and private sectors must collaborate, combining research incentives with policies that fuel innovation and competition.

AI adoption will require major changes to the skill sets that workers need, which is the goal of this course. From universities to online platforms, education must adapt to prepare an AI-enabled workforce. Blocking tools like ChatGPT isn’t a solution — embracing and integrating them is. Upskilling existing workers and rethinking how we teach future generations is essential.

And, of course, policymakers and technologists must ensure fairness, transparency, and privacy while avoiding overly burdensome regulations that stifle innovation. The challenge lies in creating frameworks that balance responsible AI use with fostering competition across organizations of all sizes.

Conclusion

In conclusion, the integration of LLMs into the workforce represents a transformative shift across industries, with significant potential to enhance productivity, creativity, and decision-making. Their current capabilities are particularly impactful in routine, repetitive, and information-intensive tasks, while human strengths such as creativity, critical thinking, and emotional intelligence remain indispensable. As LLM technology progresses, it will likely transition from assistive roles to more autonomous functions, making it essential for workers to adapt and learn to collaborate effectively with AI.

The emergence of new roles in AI development, prompt engineering, and ethics highlights opportunities for growth in the AI-driven job landscape.

To thrive in this evolving environment, individuals and organizations must prioritize reskilling and upskilling efforts, focusing on complementary human-AI workflows. The future of work is not about humans versus AI but about forging a collaborative partnership that leverages the strengths of both. Preparing for this shift will be key to unlocking the full potential of LLMs in the years to come. And we are here to help you do that!