Post-Covid, employers in the world’s largest economies over-hired. As the pandemic wound down, OpenAI’s ChatGPT made its debut in December 2023, kicking off the Large Language Model (LLM) race. It immediately ignited global interest in computer science and, for a while, university degree programs in other technical disciplines (notably, engineering) struggled to recruit. Even here at home, some universities had to make the decision to shutter entire departments because they attracted no applications.
As big-tech competitors began releasing successive iterations of their LLMs every few months and the scale of investment in capital expenditure and talent became visible, investors spent much of 2025 asking questions about where the return on those investments would come from. Tech companies began deploying LLMs internally to write (more) code and do it faster, to claims of great productivity gains, thereby cannibalizing the software industry.
But what would be the must-have killer app that would pay for it all? How does the ability to generate text, image, and video translate into greater productivity for everyone else? In the meantime, recruitment at many employers entered into a holding pattern: ‘Let’s wait and see what comes out of this AI thing and figure out if it is possible to make it do the work of entry-level positions.’
Last year saw the introduction of a new term to the AI discourse that gave businesses some hope of answering that question: Agentic AI (or AI agents). According to one definition, an AI agent is a software programme that acts autonomously to achieve a specific goal by planning, using tools and taking actions, rather than just generating text. Give an AI agent access to your email, credit card, and other accounts and credentials, and you can give it a lot of the same tasks you would give to a human personal assistant to complete on your behalf. In a sense, these AI agents are applications built on LLMs.
That makes this a tricky time for high-school students trying to decide which university programme to apply to. Is their intended profession at risk of worker displacement by AI? More urgently perhaps, what are recent graduates to do who are entering the workplace just as many businesses have hit the pause button on recruitment? As a parent of a child at this stage in their life, these are some of the questions that have kept me occupied lately.
One thing I had to remind myself of, and make peace with, is that now more than ever, careers rarely develop as linearly as planned. Life can take it to unexpected, unplanned and often better places. If you are like a lot of people and are taking a non-linear path to a career, do not discount the journey you are taking, and do not discount work you may be doing for which you consider yourself overqualified. Linear career progression is becoming an increasing rarity. Now more than at any other time, do not compare the path you are taking to that of others.
I myself zigged and zagged through several very different jobs before I found something that took me on a path that is more than a paycheck to me. One of those jobs was in the call centre of the Office of Survey Research of my university, which I worked for about a year before beginning my grad school journey. The office conducted contracted survey work. At the time, it felt like a definitive step down, but the work allowed me to develop interviewing skills and taught me how to quickly build rapport with strangers, an ability that has been indispensable in my later work in education research and the development sector.
Many jobs (and by ‘jobs’ I mean all work – working for others as well as self-employed entrepreneurs) that do not require a university degree teach transferable, lifelong, and universally applicable skills. For example, frontline customer service positions teach workers how to handle pressure, empathy, emotional intelligence, and conflict resolution. Sales positions help hone persuasion, negotiation, goal setting, performance tracking, presentation and confidence. Clerical positions are great to pick up organisation, time management and professional communication skills. At the very least, such jobs allow you to develop people skills that are needed in all workplaces.
A lot of advice these days centres on building a ‘personal brand’, but as I wade through the morass of AI slop that is my LinkedIn feed, I feel that many people simply take that to mean redirecting their humble brags, political venting, engagement farming and even thirst traps from social media platforms to LinkedIn. At the same time, several people I speak with who frequently hire new graduates are wary, not of knowledge gaps, but widespread lack of trustworthiness and consistency. People who promised the moon before being hired do not take ownership of assigned tasks and exhibit no initiative, even when all that stands between them and completing a task is a mundane Google search.
Develop a reputation for professionalism and integrity. Be known as someone who relentlessly honors their commitments. When AI can perform many functions, your human reliability and professional consistency become the irreplaceable traits that separate you and lead to future opportunities, often through word-of-mouth recommendations, within your professional network.
Cultivate a diverse set of interests and lean into their communities. Everyone should be able to lead a conversation about something other than cricket, politics, food, or work. It will make you a more interesting person and connect you with more communities – think of it as a less blatantly obvious way to network with people.
What can that look like? A recent graduate in Biology and Biochemistry from Pakistan, whom I know, could not immediately find the kind of job she was qualified for. She is also making up her mind about what graduate programme she wants to go into some years down the road. So, in the meantime, she is volunteering at a hospital’s hospice center, assisting nurses while collecting hours of volunteer work necessary for some MS programmes in the Life Sciences. She also leveraged her personal interest in fitness and workouts to earn a professional certification in personal training and nutrition to not only help her qualify for more opportunities, but also because it is relevant to one of her grad school interests (kinesiology).
Finally, she found a part-time sales position with a luxury cosmetics company where she is developing all the people skills that working in sales brings with it and covers her living expenses. It is not difficult to imagine how some or all of these efforts, skills and competencies can come together to open doors to more opportunities. And yes, beyond a certain point, you will just have to trust fate and faith, as it always has been.
Finally, stay curious. New technologies have always imposed changing workplace readiness expectations on new graduates. It was note-taking and typing in the ‘70s and ‘80s, computer programming and workplace productivity tools in the ‘90s, internet use and web authoring in the 2000s and app development in the 2010s. Today, the equivalent expectation is fluency with AI tools, bringing me full circle to where I started. Today’s workplaces demand a lot more dynamism from workers. The days when people in a 35-year work-life could plan to work in hands-on entry-level positions for no longer than five or at most ten years and then hope to graduate into a supervisory role, in which they would spend the rest of their lives pushing papers, are long gone.
The bottom-line is, Autonomous AI is not a death sentence for human labour, but a call to elevate the irreplaceable human elements of work: adaptability, integrity and genuine connection. Look beyond traditional career ladders and view every role as a platform to hone these non-negotiable qualities.
The writer (she/her) has a PhD in Education.
