AI is getting absurdly good at the math.
It can summarize a 40 page report in 12 seconds. It can spot patterns in customer behavior that would take a team a week. It can write code, rewrite it, test it, and then politely tell you your variable names are confusing.
So yeah. A lot of people feel that low hum of panic.
But here’s the calmer way to look at it.
When machines handle more of the calculation, humans don’t become useless. Humans become responsible for something else.
Meaning. And connection.
In practice, that turns into two skills that keep showing up in every “future of work” conversation that’s actually worth having.
- Data literacy (the meaning)
- Human empathy (the connection)
Not “learn to prompt better.” Not “become a 10x AI power user.” Those are fine. Helpful, even.
But if you want something sturdier than whatever tool is trending this month, it’s these two.
Let’s talk about them like real people.
Why “AI will do everything” is the wrong framing
AI can generate output. Tons of it.
But output is not impact.
You can ask a model for a dashboard summary and get a neat paragraph back. You can ask it for “top reasons churn is rising” and it will give you a list that sounds plausible. Sometimes it’s even right.
The risk is that we start treating “sounds right” as “is right.”
And even when it is right. There’s another gap.
AI can tell you what happened. It can guess why. It can propose what to do next.
But it cannot decide what matters.
That part is on us.
Meaning is not computed. It’s chosen. It’s argued for. It’s tied to context. It’s tied to values. And usually, it’s tied to people.
Which is why the human side of tech is about getting clearer, not louder.
Skill #1: Data literacy (meaning)
Data literacy is not “being good at Excel.”
It’s the ability to look at a number and ask, calmly and consistently:
What does this actually mean… and what does it not mean?
It’s the difference between reporting metrics and understanding reality.
And AI, ironically, makes this more important, not less. Because when analysis becomes cheap and instant, the world gets flooded with charts, summaries, and confident conclusions.
You need the skill to not drown in that.
What data literacy looks like in real life
Data literate people do a few things really well. Not in a show off way. More like a quiet superpower.
They ask:
- Where did this data come from? Who collected it. How. Under what incentives.
- What’s missing? What’s not measured, or what can’t be measured.
- What changed? New tracking, new pricing, a product change, a seasonality spike, a one time promo.
- What’s the definition? “Active user” can mean 10 different things across teams. Same with “conversion.”
- What’s the base rate? A 30% increase on a tiny number can be meaningless. Or it can be early signal. You have to check.
- What decision is this for? If the decision is unclear, the analysis will wander forever.
And they’re comfortable saying a sentence that is weirdly rare in workplaces:
“I don’t know yet. We need one more check before we act on this.”
That sentence saves companies millions. It also saves a lot of careers, quietly.
AI can do the analysis, but you have to do the interpretation
Here’s an example that happens constantly.
AI says: “Churn increased due to poor onboarding.”
Maybe. Could be true.
But a data literate human immediately asks:
- Did churn increase for all cohorts, or only new users?
- Did churn increase right after signup, or after week 3?
- Did anything change in the product, pricing, marketing channel mix?
- Are we mixing free users and paid users?
- Are we sure the churn metric is calculated the same way as last month?
This is the meaning work.
AI can propose explanations. But humans need to test them against context. Against definitions. Against how the business actually runs.
Data literacy is a leadership skill now
It used to be acceptable to outsource “numbers” to the analyst team and just wait for the slides.
That era is ending.
Because now everyone can generate slides.
So the real question becomes: who can tell the difference between a compelling story and a correct one?
Data literacy is how you keep AI from becoming a confidence machine. It helps you turn data into decisions instead of decoration.
And yes, you can learn this. You don’t need a stats degree. You just need reps.
Small reps count.
Pick one metric at work and learn its definition. Where it’s stored. What makes it move. What makes it lie.
That alone puts you ahead of most rooms.
Skill #2: Human empathy (connection)
Empathy gets misunderstood in tech.
People hear it and think it means being soft. Or being agreeable. Or sending more smiley faces.
No.
Empathy is a precision skill. It’s the ability to understand what someone is experiencing, and respond in a way that actually helps.
It’s connection, yes. But it’s also clarity. Boundaries. Timing. Translation.
And AI can mimic empathy. It can generate “I’m sorry you’re going through that” in 200 styles.
But mimicry isn’t the same thing as relationship.
Humans can feel when they’re being handled.
What human empathy looks like at work
Empathy is:
- The PM who notices support is overwhelmed and changes the rollout plan.
- The engineer who explains tradeoffs without making the other person feel stupid.
- The manager who gives feedback in a way that doesn’t collapse someone’s confidence.
- The designer who sits with a user’s frustration long enough to find the real problem.
- The analyst who doesn’t just drop a scary metric in Slack and disappear.
Empathy is also knowing that people don’t resist change because they’re irrational.
They resist change because change costs something.
Time. Status. Safety. Identity. Competence. Predictability.
If you can sense those costs, you can lead better. You can design better. You can sell better. You can support customers better.
AI can’t do that in a fully real way because the core ingredient is trust. And trust is earned in relationship, over time, with consequences.
The “connection” part is where products win or lose
Most products don’t fail because the model wasn’t accurate enough.
They fail because people didn’t adopt them. Or they churned. Or they never understood the value. Or they felt ignored when something broke.
Empathy is what keeps the loop tight between what you build and what humans can actually use.
It’s also what keeps teams from turning into robots who ship features while slowly resenting each other.
AI can help write the message. But it can’t feel the moment.
You can.
The real shift: AI handles the math, humans handle meaning and connection
This is the part I want to land cleanly.
The future is not “humans vs AI.”
It’s more like:
- AI accelerates the work.
- Humans decide what the work is for.
Meaning and connection are the two anchors that keep acceleration from turning into chaos.
Data literacy stops you from making confident decisions on shaky ground.
Empathy stops you from treating people like inputs and outputs.
And if you combine them, you get something rare.
You become the person who can translate between numbers and humans.
That’s basically the job now. In product. In marketing. In ops. In engineering. In leadership. Even in individual contributor roles, honestly.
The person who can say:
“Here’s what the data suggests, here’s how reliable it is, and here’s what it will feel like for the customer and the team if we act on it.”
That person is hard to replace.
How to build these skills without making it a whole personality
A simple, realistic approach. No reinvention required.
To build data literacy
- Pick one metric you touch weekly. Learn its definition like it’s your job. Because it kind of is.
- Ask “compared to what?” every time you see a number. Last week, last month, same period last year, control group, baseline.
- Look for one alternative explanation before you accept the first story.
- Write down assumptions in plain English. If they feel embarrassing, good. Now you can test them.
To build empathy
- Practice reflecting. “What I’m hearing is…” and then say it back. This alone reduces conflict by a lot.
- Ask one more question than you think you need. Especially when someone is upset or vague.
- Separate intent from impact. People often mean well and still cause harm. Naming that gently is a skill.
- Follow through. Empathy without follow through feels fake. Even a small update builds trust.
None of this requires you to be extroverted. Or “a people person.” It’s not personality.
It’s craftsmanship.
A small, hopeful conclusion
AI is going to keep getting better at the math. Faster, cheaper, more automatic.
Good.
Let it.
Your edge is not competing with a machine on speed. Your edge is doing what machines don’t own.
Choosing what matters. And caring how it lands.
Data literacy gives you the meaning.
Human empathy gives you the connection.
And if you lean into both, you don’t just stay relevant. You become the reason the tech works at all.
FAQs (Frequently Asked Questions)
Why is the framing ‘AI will do everything’ considered wrong?
The idea that ‘AI will do everything’ is misleading because while AI can generate vast amounts of output and even propose explanations, it cannot decide what truly matters. Meaning is not computed; it’s chosen, argued for, and tied to context, values, and people. Humans are essential for interpreting AI outputs to create real impact.
What is data literacy and why is it important in the age of AI?
Data literacy is the ability to look at numbers calmly and consistently ask what they actually mean and what they do not mean. In an era where AI makes analysis cheap and instant, data literacy helps individuals avoid drowning in charts and confident conclusions by critically evaluating data sources, definitions, changes, and context to make informed decisions.
How does AI impact the role of humans in interpreting data?
AI can perform analysis quickly and generate plausible explanations, but humans must interpret these results by questioning assumptions, understanding context, verifying definitions, and testing hypotheses against how the business operates. This human interpretation ensures that decisions are based on meaning rather than just data output.
Why is human empathy considered a critical skill alongside data literacy?
Human empathy is a precision skill involving understanding others’ experiences and responding helpfully. Unlike AI’s mimicry of empathy through generated responses, genuine empathy creates real connection through clarity, boundaries, timing, and translation—skills essential for effective teamwork, leadership, and communication in tech environments.
What practical behaviors demonstrate data literacy in the workplace?
Data literate individuals routinely ask about data origins, what might be missing or unmeasured, changes affecting metrics (like seasonality or product updates), precise definitions of terms like ‘active user,’ base rates behind percentages, and the decision context for analysis. They are also comfortable admitting uncertainty when more checks are needed before acting.
How can one develop stronger skills in data literacy and human empathy?
Developing these skills involves consistent practice: for data literacy, start by deeply understanding one metric at work—its definition, storage location, factors influencing it—and question its meaning regularly. For human empathy, practice active listening, clear communication with respect for others’ perspectives and feelings, setting boundaries thoughtfully, and responding in ways that genuinely support colleagues.
