We all know it. AI is coming for us.
It’s no longer a matter of "if" but "when". When will AI come after our jobs and nudge us off the picture?
No one seems to be safe. AI threatens to replace almost all jobs from all industries. Whether it be through robotics, task delegation or automations, we all saw AI performing surgeries, creating motion pictures and more.
And it’s only the beginning. Can’t even imagine what it’ll be capable of in a few years, better be prepared, right?
But how? How do you plan for the unknown? How do you navigate a career when there is a risk of getting fully replaced?
This question has been sitting rent-free in my mind lately, and I’ve reflected a lot on it.
The solution is figuring out how much of your role AI will claim and which parts will still value human input
What came out of this reflection was surprisingly positive because I no longer feel scared. I’ve tamed those fears and I’m here to help you achieve the same.
At the end, I’ve managed to devise a survival strategy to make sure both I and AI keep our places in the same picture. Let’s do the same for you!
To figure out what moves you need to make to ensure you remain as relevant as possible in the job market, you need to:
- Accept AI rather than live in fear. I’ll show with you how I did it.
- Analyse the essence of your job. I’ll help you reflect on its core tasks.
- Identify what AI can’t replace. Your survival depends on defining these elements accurately.
BIG News – I’ve Started My Newsletter!!!
I know lots of you are curious about what it’s like to work in one of the most adored companies out there. So I’m taking you backstage!
Each week, I’ll share one quick, digestible story and lessons from a day in my life as a Spotify data scientist.
Make sure to subscribe 👇 It’s free
1. Stop Freaking Out About AI Replacing You
Just because AI is capable of doing many of the tasks related to your job doesn’t mean it’ll replace you. I’m not the one saying it, history is.
Humans have gone through uncountable technological shifts in the past that claimed many jobs. Yet many also survived these shifts. What happened is that most of these changes replaced jobs "as we know them" at a given point in time.
It’s not the jobs themselves that vanish into thin air; it’s the tasks within them that get a makeover.
To stop worrying about AI replacing your job, you first need to understand what a job means in the first place. I bet you haven’t really thought about it.
Me neither.
What’s a job anyway?
A job is a bundle of tasks or skills, all knitted together towards a common goal. And what usually happens when disruptive tech waltzes in is that it does not erase the goal. It automates the steps we take to reach it.
So the question stops being about whether your job will be replaced or not, but the extent to which it will be.
This shifts the narrative from "Will my job survive?" to "How will my job evolve?"
For us data scientists – and really, for anyone worrying about their career – it’s about peeling back the layers to understand the essence of our work.
What does being a data scientist really mean?
And if you’re not in Data Science, this introspection is just as crucial for your field.
I kind of already answered this question in this article:
Don’t Apply to Tech Without Mastering These 6 Must-Have Data Science Skills – A Spotify Data…
Let’s break down those skills and directly ask an AI how it’s already automating them. I’ll use ChatGPT in this case.
We’ll then try to figure out what aspects of those skills you’ll have to focus on to avoid getting kicked out of the frame.
DS Skill #1. Extracting and preparing data from various sources. This involves cleaning, transforming and shaping the raw data.
"AI can assist with this task, especially in generating SQL queries based on natural language descriptions or optimising existing queries.
However, the nuanced cleaning, transforming, and shaping of raw data often require human judgment to understand the context and specific needs of a project."
DS Skill #2. Exploring and analysing the prepared data using statistical and visualisation techniques to understand trends, outliers and patterns.
"AI can perform exploratory data analysis to some extent. This includes identifying trends, outliers, and patterns in data using statistical techniques.
AI can also generate descriptive statistics and insights from data, although nuanced interpretation and understanding of complex patterns might still benefit from human expertise."
DS Skill #3. Designing visualisations like charts, graphs and dashboards to tell stories with the data and communicate findings to stakeholders.
"AI can suggest or even generate basic visualisations and dashboards based on data.
However, designing comprehensive and insightful visualisations that tell a compelling story with the data often requires human creativity and an understanding of the audience’s needs."
DS Skill #4. Building predictive models using techniques like machine learning and A/B testing to forecast outcomes and inform decisions.
"AI is at the core of building predictive models. AI systems can select appropriate algorithms, train models on datasets, and evaluate their performance using techniques like cross-validation.
However, choosing the right model architecture, feature engineering, and interpreting model outcomes in a business context usually benefit from human expertise."
DS Skill #5. Monitoring key performance indicators (KPIs) and metrics to measure business performance and the impact of decisions.
"AI can monitor KPIs and metrics effectively, especially with the integration of automated systems that track and report these indicators in real-time. AI can also identify anomalies or significant changes in performance metrics, alerting humans to potential issues.
However, understanding the root causes behind changes in KPIs and making strategic decisions based on this information often require human insight."
Focusing on human-centric skills
I bet you can see a pattern already. Soon AI will be able to assist with and/or perform many tasks within the data science workflow, particularly those involving data processing, analysis, and modelling.
What won’t be replaced are tasks that require deep contextual understanding, strategic decision-making, and creative storytelling.
Data scientists won’t be replaced but we’re likely to see a big shift in the nature of how we perform our work.
Instead of coding, AI might do it in our stead.
But the twist here is that we’ll be responsible for making sure it does it well, and for tying everything together across all business goals and related projects.
Instead of getting replaced, we’re likely to enter a collaborative partnership where we complement each other’s abilities.
However, becoming complimentary to AI means that we’ll have to shift our focus to honing the skills where we surpass AI – where human touch will always prevail.
2. Survival of The Fittest
We all know computers as they are today, they’re portable and anyone can use them. But did you know computers used to be humans less than a century ago?
Actual people used to do all the heavy computing lifting by hand.
But when IBM introduced the first ever computer, those who didn’t lose their jobs were those who saw it coming and quickly adapted to it.
They stopped being computers and instead became computer scientists.
The movie Hidden Figures tells a beautiful story of how the women computers in NASA managed to quickly pivot to being computer scientists when they were threatened to lose their jobs. I highly recommend it.
So the "fittest" here becomes the most adaptable – the ones that readily adjust to different environmental conditions.
Adaptability is the greatest skill for when life throws at you curveballs, and you don’t know how to take them.
You actually don’t have to know how to take them, you just need to be ready for them, and for that, you need to keep your eyes wide open.
How to join the adaptive species?
-
Stay aware to stay afloat. It’s too easy to disconnect with trends once you’re settled in your job. So I make it a conscious effort to keep track of tech advancements regardless of whether I care or not. That way, I’m sure I’ll adapt and pivot when needed. TDS is definitely a blessing for the DS community; it brings us the latest of what’s going on in the field. So find your own publication, journal or community.
-
Diversify your skillset. Should I go full stack data science and develop ML engineering skills? or should I focus more on developing soft skills? The truth is it might be safer to become a Jack of all trades. Don’t limit yourself to your current expertise. Mitigate the risk by exploring adjacent areas like ML engineering, data governance, or non-technical skills like project management.
- Anticipate industry shifts. I’m constantly analysing arising trends in the background to predict future changes in the field. It helps me reflect on how these changes might impact my role. That way, I prepare accordingly. This whole article is a summary of my analysis.
-
Remain open to new opportunities. Whether it’s a new project, role, or learning a new tech, embrace opportunities that push you out of your comfort zone. Don’t be afraid to try new things and fail. The faster you fail, the quicker you learn what works and what doesn’t, which is a crucial rule in being adaptable in life.
- Assess the evolving demands of the market. Keep an eye on which skills are in demand and which are becoming obsolete. Check out some job posts on LinkedIn and assess what skills employers are seeking. This can help you decide whether you need to deepen your expertise in a niche area or broaden your skill set.
Follow Jack, he knows da way
I’d like to think that if you can dip your toes in all sorts of different sauces, then you might be playing smarter than someone betting everything on the BBQ dip. I personally can never seem to pick only one, I prefer mixing things.
My bet? Becoming a Jack of all trades. I can’t afford to put all my eggs into the same basket.
So when in doubt, I apply the rule of portfolio diversification. It’s the golden rule for risk mitigation.
I recently asked a friend who’s a freelance data scientist what skills were the most valued in the market.
Surprisingly, she said "getting things done".
The ability to take on any data science project and quickly develop the skills when needed is the master dip.
Knowing a bit of many things means you can rapidly shift from one skill to another based on the demands of the market. But to achieve that, you need to master the fundamentals that help you play with different aspects of data science.
3. To Business or Not To Business
Let’s bounce back on what ChatGPT told us earlier:
"Nuanced cleaning, transforming, and shaping of raw data often require human judgment to understand the context and specific needs of a project.
"Nuanced interpretation and understanding of complex patterns might still benefit from human expertise."
"Designing comprehensive visualisations that tell a compelling story with the data often requires human creativity and an understanding of the audience’s needs."
"Understanding the root causes behind changes in KPIs and making strategic decisions based on this information often require human insight."
"Choosing the right model architecture, feature engineering, and interpreting model outcomes in a business context usually benefit from human expertise."
You get the picture.
Understanding the business context and strategic implications of data insights is something AI cannot fully replicate.
I mean even AI knows it.
So you need to focus on learning how your work impacts the business you work in. You need to become a key business player.
Here’s how to develop business acumen in your company
- Seek cross-functional collaboration. Look for opportunities to work with different departments or stakeholders. Understand their challenges and goals, and you’ll learn valuable business insights.
- Participate in business strategy meetings. If possible, get involved in strategy meetings or project kick-offs to gain a broader perspective on business priorities and objectives.
- Learn basic finance and economics. Familiarise yourself with fundamental financial concepts, such as profit and loss, return on investment, and cost-benefit analysis to better understand the financial implications of your projects. A company is a business after all.
- Practice developing forecasts and scenarios based on data insights. Consider how different strategies could affect outcomes, both in the short and long term.
- Focus on problem-solving through business lenses. Approach data science problems with a focus on how your solutions can drive business value, such as increasing revenue, reducing costs, or enhancing customer/user satisfaction.
Domain expertise keeping us safe and warm
As data scientists, our work is highly dependent on something called domain expertise, which in most cases lives in people’s minds.
So unless we start documenting literally everything and feed that to an AI continuously, we’re relatively safe from the threat.
AI is for sure capable of pattern recognition. It’s the 101 of what we learn in ML, but humans will always be needed to make sure those interpretations and recommendations match reality. And to do that, we’ll definitely need our domain and technical expertise to review the output produced by AI.
4. Softening Things Up
Business acumen isn’t the only thing ChatGPT insisted on.
AI will definitely do the job, but it won’t be the one delivering it to humans. For that it needs another human, but not just any. Someone who masters the fundamentals of the job – someone like you.
Effective communication bridges the gap between data science and business impact.
It makes you an essential liaison capable of translating data into strategic insights that resonate with stakeholders.
One thing that makes us human is that despite all our daily interactions with Technology, we place human touch at the highest regard. It will most likely keep being so in a tomorrow where machines will be ever more ubiquitous.
This means that you’ll have to draw your focus on developing your soft skills more than the hard ones.
Here’s how to develop your communication, storytelling, and stakeholder management skills
- Learn the principles of storytelling. Study storytelling techniques and understand how to structure a narrative that captivates and informs your audience. I recently purchased storytelling cards, but you can always find plenty of other resources online.
- Enhance your skills in data visualisation tools and methods. Learn how to use visual elements to tell a story and highlight key findings. I like this chart suggestion guide, it helps understand how and when to use the different charts out there.
- Practice public speaking to different audiences. Use feedback from your peers to refine your delivery and make your presentations more engaging and impactful. Make yourself noticed and you’ll become more valuable.
- Manage stakeholder expectations. Through clear and honest communication, setting realistic timelines, and delivering on your promises.
- Cultivate active listening. Understand stakeholder feedback and concerns fully. This can guide you in providing more targeted and useful insights, and will also make you a valuable team player.
5. Making Friends With AI
Have you ever heard the expression "keep your friends close and your ennemies closer?" This also extends to our all-conquering AI frienemy.
If you can’t fight it, then better make friends with it, right?
But even that won’t be enough on its own. You must **** complement AI mastery with a solid foundational data science core.
Here’s how to augment your AI know-how
- Strengthen statistical foundations. Once AI starts automating technical tasks, we’ll still be expected to troubleshoot, monitor and interpret results to non-technical stakeholders. We’ll still act as a bridge, and it means mastering the underlying mathematical principles.
- Build a solid understanding of AI mechanics. Dive deeper into how AI and ML models work, including neural networks, reinforcement learning, natural language processing techniques and LLMs. This will help you become a responsible AI practitioner and leader – the data scientist of tomorrow.
- Enhance your troubleshooting abilities. Develop a keen eye for diagnosing and troubleshooting AI model issues, from data discrepancies to model biases and errors.
- Learn the art of prompting. Understand how to craft effective prompts for AI models. This includes being concise yet detailed, understanding the model’s capabilities, and iterating on prompts based on outcomes. In a word, learn to speak to AI.
- Automate routine tasks with AI. Identify opportunities within your daily tasks where AI can automate processes or assist such as coding or data analysis. This will free up your time for more strategic work.
AI won’t replace you but someone who writes better prompts will.
My friend Nabil Alouani is a great teacher on how to become an Ace at prompt engineering. Be sure to check out his articles! He also crafted the most informative and comprehensive guide on how to upskill your prompt engineering game.
How to Write Expert Prompts for ChatGPT (GPT-4) and Other Language Models
Final Word
Ultimately, I don’t think AI will replace us data scientists, and I hope my perspective isn’t just a product of wishful thinking. I also believe it is safe for us to view AI under a positive light without feeling ever-so threatened by its glow.
Our light won’t fade, we’ll just glow differently, but only if we learn to adapt to AI advancements. Only then will we be able to keep our jobs.
AI will surely automate some of our tasks, but in doing that, it will also enable us to become more efficient.
Unfortunately, the same way not all human computers survived modern computers, not all data scientists will survive AI. History always repeats itself. Better come to terms with this reality now before it’s too late.
Survival will favor those who not only grasp the mathematical foundations and the nuances of prompt engineering but also possess robust business insight, superior soft skills, and the ability to act as a bridge with AI.
That’s why it’s crucial now more than ever to sit on the question and devise your own survival guide. How can you prepare?
- Tame the fear of AI displacement. Figure out what skills make up your role and identify the aspects that value human touch.
- Learn adaptability. It’s the only way survive an environment subject to the ever-changing influence of AI. Keep track of industrial changes, stay open to emerging opportunities and diversify your skill set.
- Sharpen business acumen. What won’t be replaced are tasks that require deep contextual understanding, strategic decision-making, and creative storytelling.
- Elevate soft skills . Learn the principles of storytelling, improve your public speaking and delve into stakeholder management.
- Augment AI mastery and technical core. Strengthen your statistical foundations and build a solid understanding of AI mechanics. Learn the art of prompting; the data scientist of tomorrow will master the power of speaking to AI. So should you!
I have GIFTS for you 🎁 !
Sign up to my newsletter K’s DataLadder and you’ll automatically get my ultimate SQL cheat sheet with all the queries I use every day in my job in big tech + another secret gift!
I share each week what it’s like to be a Data Scientist in Tech, alongside practical tips, skills, and stories all meant to help you level up – because no one really knows until they’re in it!
If you haven’t done that already
- Subscribe to my YouTube channel. New video coming up very soon!
- Follow me ** on Instagram, LinkedIn, ** X, whatever works for you
See you soon!
