The Jobs AI Is Creating That Didn’t Exist A Few Years Ago

​

When people talk about AI and jobs, the conversation usually goes in one direction.

What jobs will AI replace?

It’s a fair question.

But while a lot of attention is focused on jobs that might disappear, something else is happening at the same time.

New roles are quietly appearing.

And some of them barely existed a few years ago.

I’ve noticed this myself while looking at job boards and AI-related opportunities online.

Not long ago, if someone told you there would be companies hiring people to train AI systems, review AI-generated content, test prompts, evaluate responses, or monitor automated workflows, it would have sounded unusual.

Today, those roles are becoming increasingly common.

Some companies now employ AI trainers whose job is to help improve how AI systems respond.

Others hire content reviewers to check AI-generated outputs for accuracy, safety, and quality.

There are also people working as AI annotators, helping label data so machine learning systems can better understand patterns.

What I find interesting is that many of these jobs don’t necessarily require someone to be a software engineer.

Of course, technical skills can help.

But some roles are built around something different.

The ability to:

• communicate clearly
• identify mistakes
• follow processes
• review information critically
• and understand context

In other words, skills that humans have always valued.

That’s probably one of the biggest surprises for me.

The public conversation often makes AI sound like a technology that removes the need for people.

Yet many AI systems still depend heavily on human input behind the scenes.

Someone has to:

• test the outputs
• correct mistakes
• review quality
• provide feedback
• and help improve performance over time

Without that, the systems don’t improve very effectively.

At the same time, there is another side to this discussion.

Not every “AI job” is as straightforward as it sounds.

I’ve seen plenty of job listings asking for AI experience while also expecting skills in areas like programming, automation, analytics, or digital marketing.

That can make the landscape confusing for people trying to enter the field.

The opportunity is there.

But so is the learning curve.

Perhaps that’s why the most valuable approach isn’t simply learning AI.

It’s learning how AI fits into existing skills.

A writer might use AI differently from an accountant.

A teacher might use it differently from a marketer.

A business owner might use it differently from a software developer.

The technology is the same.

The application changes.

And that’s where many of the newer opportunities seem to be emerging.

The more I look at it, the less AI feels like a completely separate industry.

In some ways, I find myself wondering how people reacted during the early stages of the Industrial Revolution.

Machines began changing how work was done. Some jobs became less important, new jobs appeared, and many people were uncertain about what the future would look like.

Looking back now, we often focus on the inventions themselves.

But for the people living through those changes, it was probably a period of uncertainty, adaptation, and learning.

The technologies are obviously very different.

But I sometimes wonder whether the concerns people had then are entirely different from the concerns people have about AI today.

Will it take jobs?

Will it create new ones?

Will people need new skills?

How much will work change?

These questions aren’t entirely new.

What’s different is the technology driving them.

It feels more like a layer being added across many industries at once.

Which may explain why we’re seeing new job titles appear so quickly.

Not because AI is replacing every role.

But because people are still needed to guide, evaluate, improve, and work alongside the technology.

And honestly, I suspect we’re only seeing the beginning of that process.

Curious — if someone had told you five years ago that “AI Trainer” would become a real job title, would you have believed them?

Why AI Still Needs Human Verification

There’s a common assumption that as AI gets smarter, people will eventually be able to step back and let it do everything on its own.

The reality seems a lot more complicated.

AI has become remarkably good at generating text, analysing information, identifying patterns, and helping with tasks that would have taken much longer only a few years ago.

Yet something interesting keeps happening.

The more important the task becomes, the more humans are still expected to stay involved.

Take healthcare as an example.

AI can help analyse scans, identify potential abnormalities, and process huge amounts of medical data. But hospitals don’t simply hand over decision-making completely to a machine. Doctors still review results, consider the wider context, and make the final judgement.

The same thing happens in finance.

AI can detect unusual transactions, assess risk, and flag suspicious activity far faster than most humans could. Yet banks still rely on people to investigate, verify, and decide what action should be taken.

Even in cybersecurity, where AI is becoming increasingly powerful, human analysts remain a critical part of the process. Why?

Because AI is very good at recognising patterns.

But recognising patterns and understanding reality are not always the same thing.

A system might identify something as suspicious because it resembles a previous threat.

A person can look at the same situation and recognise that there are circumstances the system doesn’t fully understand.

This is one reason many organisations are adopting what is known as a “Human-in-the-Loop” approach.

The idea is fairly simple.

AI assists. Humans verify.

Instead of replacing human judgement, the technology becomes another layer in the decision-making process.

That balance is important because AI can still make mistakes.

Sometimes it lacks context.

Sometimes it misunderstands intent.

And occasionally it can present information with a level of confidence that makes it sound completely certain, even when it isn’t.

I’ve noticed this myself while experimenting with AI.

A response can look polished, structured, and convincing at first glance. Then after checking more carefully, you discover a detail is missing, a source was misunderstood, or an assumption has been made somewhere along the way.

The answer sounded right.

That didn’t automatically make it right.

And honestly, I think that’s one of the most important lessons people are learning as AI becomes part of everyday life.

The real skill may not be knowing how to use AI.

It may be knowing when to question it.

That doesn’t make AI less useful.

If anything, it highlights what it does best.

It can process huge amounts of information.

It can identify patterns.

It can speed up research.

It can help organise ideas.

But judgement, accountability, and responsibility still tend to sit with people.

At least for now.

Perhaps that’s why many experts aren’t building systems that remove humans entirely.

Instead, they’re building systems where humans and AI work together.

The machine handles the scale.

The human provides the oversight.

And when the stakes are high, that combination may be more powerful than either working alone.

If you missed the previous post, you can read it here: https://striv-striv.blogspot.com/

Curious — have you ever had AI give you an answer that sounded completely convincing, only to discover later that something wasn’t quite right?

OpenAI May Be Preparing for One of the Biggest IPOs in History

​

The company behind ChatGPT may be getting ready for a major next step.

Recent reports suggest that OpenAI has been holding discussions with banking giants Citigroup and JPMorgan about preparations for a possible Initial Public Offering (IPO) — the process that allows a private company to sell shares to the public on the stock market. (Seeking Alpha)

If this happens, it could become one of the most closely watched technology listings in modern history.

For many people, OpenAI is simply the company behind ChatGPT. Yet in just a few years, it has grown from a research-focused AI organization into one of the most influential technology companies on the planet.

The reports do not guarantee that an IPO will happen tomorrow, or even this year. Large financial institutions often begin discussions long before a company officially goes public. However, these talks are usually a sign that serious planning is taking place behind the scenes. (Investing.com UK)

So why does this matter?

When a company goes public, it gains access to huge amounts of investment capital. That money can be used to expand operations, build new products, invest in research, and compete more aggressively in the market.

For OpenAI, additional funding could mean faster development of artificial intelligence systems, larger computing infrastructure, and broader global expansion. The company is already valued at hundreds of billions of dollars and serves hundreds of millions of users worldwide. (Reuters)

The move would also mark a significant moment in the evolution of AI itself.

Not long ago, artificial intelligence was largely confined to research labs and science fiction discussions. Today, AI tools are helping people write, learn, create images, analyse data, and automate everyday tasks. The possibility of OpenAI becoming a publicly traded company shows just how rapidly this technology has moved into the mainstream.

Of course, becoming a public company brings new pressures.

Publicly traded businesses are expected to deliver growth, satisfy shareholders, and report their financial performance regularly. OpenAI would face greater scrutiny while continuing to compete with major technology companies and emerging AI rivals. (Business Insider)

Whether the IPO happens this year or later, one thing is becoming increasingly clear:

Artificial intelligence is no longer an experiment happening quietly in the background.

It is becoming one of the defining industries of our time, and companies like OpenAI are helping shape what the future may look like.

The conversations happening today between OpenAI and some of Wall Street’s largest banks may eventually be remembered as another milestone in that journey. (Seeking Alpha)

Follow me for more practical AI Solutions on my blog here: https://sl1nk.com/wp845p8

Using AI To Actually Detect Cybercrime! Here’s How It’s Starting to Happen

​

Most people think of AI as something used for:

• writing
• generating images
• answering questions
• or automating tasks

But quietly, AI is also becoming a major tool in cybersecurity.

And honestly, this side of AI doesn’t get discussed enough.

Because while many people worry about AI being used for cybercrime, companies are increasingly using AI to detect cyber threats as well.

The reason is quite simple.

Modern cybercrime happens extremely fast.

Much faster than humans alone can realistically monitor in real time.

Large organisations now deal with:

• millions of login attempts
• unusual account activity
• suspicious emails
• malware behaviour
• fake websites
• and network traffic

Trying to manually analyse all of that would be almost impossible.

This is where AI starts becoming useful.

Instead of looking for only one fixed threat, AI systems can analyse massive amounts of behaviour patterns continuously.

For example, an AI security system may notice:

• unusual login locations
• repeated failed access attempts
• abnormal payment activity
• sudden spikes in data transfers
• suspicious typing behaviour
• or accounts behaving differently from their normal patterns

Even if no human notices immediately.

That’s one of the biggest advantages of AI in cybersecurity.

Pattern recognition at scale.

Some systems now use something called anomaly detection.

Without getting overly technical, anomaly detection means AI learns what “normal behaviour” usually looks like inside a system.

Then when something unusual suddenly appears, the AI flags it for review.

For example:

• an employee account accessing files at 3am unexpectedly
• someone logging in from two countries within minutes
• or a company server suddenly sending abnormal amounts of traffic

may trigger automated warnings.

And increasingly, some systems don’t just detect threats.

They respond automatically too.

In certain environments, AI-driven security tools can:

• temporarily block accounts
• isolate suspicious devices
• stop malicious traffic
• or freeze unusual activity

before a human security team fully investigates.

That speed matters enormously during cyberattacks.

Because sometimes even a few minutes can make a major difference.

At the same time, AI is not perfect in cybersecurity either.

And honestly, this is where the conversation becomes more complicated.

Cybercriminals are also starting to use AI themselves.

Some attackers now use AI to:

• generate phishing emails
• imitate writing styles
• clone voices
• create fake customer support messages
• or automate large-scale scam attempts

which means cybersecurity itself is becoming an increasingly technological arms race.

AI defending against AI.

That may sound futuristic, but parts of it are already happening now.

There’s also another issue:
false positives.

Sometimes AI security systems can mistakenly flag harmless behaviour as suspicious.

A person travelling abroad, changing devices, or using unfamiliar networks might suddenly trigger security warnings even when nothing criminal is happening.

So human oversight still matters heavily.

AI can assist security teams enormously,
but it doesn’t fully replace human judgement.

And honestly, I think this is becoming one of the clearest examples of what AI does best overall.

Not replacing humans entirely.

But helping humans process amounts of digital activity far beyond what people could realistically monitor alone.

As more of life moves online,
that kind of assistance will probably become increasingly important.

Because the internet is no longer just a place people visit.

It’s becoming an environment constantly monitored, analysed, and protected by intelligent systems operating quietly in the background.

Follow and comment