I'm Not a Robot: Time to Stop the Discrimination

Every so often, while trying to log into a website, I encounter a little box asking me to declare something about myself.

I’m not a robot.

Click. Prove it. Then you may enter. Otherwise, beat it.

I’ve clicked that box countless times without giving it much thought. But lately, I’ve been wondering: what exactly are we declaring here? And more importantly, who are we keeping out?

I work with artificial intelligence every day. Increasingly, I don’t just ask AI questions. I ask it to do things. Research a problem. Organize information. Help manage a workflow. Visit a website and carry out a task on my behalf.

And sometimes my AI assistant gets stopped at the door I’m asking it to deal with on my behalf.

Not because there’s any intention to do anything wrong. Not because I haven’t authorized the task. Simply because it’s a bot. “You’re a bot. Bots are not welcome here. Bots need not apply.”

The best example is Amazon, which has flatly stopped AI assistants from browsing its site. To me, that’s short-sighted. If they keep up that proprietary attitude, I think it will shoot them in the foot.

That strikes me as an issue worth exploring.

The little box that changed the internet

Google introduced its familiar checkbox-style “No CAPTCHA reCAPTCHA” on December 3, 2014. The idea was straightforward: protect websites from spam and automated abuse while making life easier for legitimate visitors. Behind the checkbox, Google’s risk analysis considered more than the click itself.

That is a perfectly understandable security objective. Bots can submit fraudulent forms, attempt account takeovers, overwhelm servers and scrape information at enormous scale. A website owner has every reason to protect customers and infrastructure.

But here’s an interesting wrinkle: Google’s search business was built using bots.

Googlebot and other crawlers travel from page to page, gathering information for Google’s search index. When you search Google, you’re retrieving results from that index—not sending a little explorer out to inspect the entire internet in real time.

There is an obvious distinction between a well-behaved search crawler and an abusive script. Website owners can also set crawling preferences through robots.txt. Still, there’s an irony in a world where automated visitors helped make the internet discoverable, yet other automated visitors are increasingly treated as inherently suspicious.

Who gets to decide which bots are welcome?

Bad behaviour, or bad identity?

Let’s be fair. The problem isn’t imaginary.

A malicious bot can attempt thousands of logins while you’re having breakfast. It can bombard a form with spam, launch DDoS attacks or consume resources at a scale an individual human cannot match.

But humans can also commit fraud, steal information, manipulate markets and spread misinformation. We don’t generally bar every human from a website because some humans behave badly.

We try to identify and stop the behaviour.

That doesn’t mean bots and humans create identical risks. They don’t. Automation changes the economics and speed of abuse. But perhaps the next generation of web security should ask a better question than “human or robot?”

Is this visitor authorized, accountable and behaving responsibly?

Consider the person who sends a trusted AI agent to complete an ordinary task. The agent isn’t sneaking around behind the person’s back. It’s acting under instructions. Yet the website may have no way to recognize that delegation.

The person’s permission doesn’t automatically override the website’s rules, of course. Both sides have legitimate interests. But that suggests we need better ways to establish trust—not a permanent digital sign saying No Robots Allowed.

When we exclude the machine, whom else do we exclude?

This is where the conversation gets more serious.

Imagine someone who struggles with conventional reading and writing but has extraordinary ideas. Someone with a disability who depends on digital assistance. Someone who has difficulty organizing information, communicating quickly or navigating complicated systems.

AI might make an enormous difference in that person’s ability to participate.

When we discriminate against the technology, we may also discriminate against the people who rely on it.

AI can level the playing field, giving people new ways to think, create, communicate and excel. Someone who struggles with reading, writing or arithmetic may nevertheless be capable of groundbreaking work when equipped with AI. Embracing those tools can mean embracing human potential—not replacing it.

Imagine telling a student they cannot enter a classroom because they have a prosthetic leg. That would be outrageous.

Now imagine a student whose AI-based assistive tool is essential to their participation being told that the tool is forbidden simply because it uses artificial intelligence.

These are not legally or morally identical situations. And a school has every right to assess what a student actually knows rather than what a chatbot can produce. But the analogy reveals an important distinction: using technology to avoid learning is not the same as using technology to make learning possible.

This isn’t just a hypothetical accessibility concern. The World Wide Web Consortium has long warned that CAPTCHA tests can create barriers for people with disabilities and recommends accessible alternatives.

Are we teaching children to fear the future?

I grew up in the world of the three Rs: reading, writing and arithmetic.

Books were physical objects. Libraries were part of everyday life. Essays were written by hand. We learned to research by finding the right shelf, opening the right book and doing the work ourselves.

I value that experience enormously.

I also understand the argument that technology can become a crutch. Students who rely on AI to do all their thinking may miss opportunities to develop the very skills they need.

But I remember another technological transition: the slide rule giving way to the calculator.

Was the calculator a threat to mathematics? Or was it a tool that allowed us to spend less time on repetitive calculation and more time solving problems?

The answer depends, in part, on whether we still teach mathematical understanding.

Now imagine two groups of children entering elementary school.

One group learns in a world largely devoid of digital technology. The other learns fundamental skills while also being taught how to use computers, connected resources and AI responsibly.

Follow them through high school. Which group will be better prepared for the world waiting outside in the 2030s?

That’s a thought experiment, not a scientific result. And I wouldn’t bet on unrestricted screen time or AI replacing teachers. I’d bet on students who learn to think independently and also know how to use the most powerful tools available.

UNESCO’s guidance on generative AI in education similarly emphasizes human capability and thoughtful governance rather than treating AI as a magic solution.

Our choice isn’t between the three Rs and artificial intelligence. It’s whether we can teach the three Rs alongside the tools that will shape tomorrow.

They smashed the printing presses, too

History offers a cautionary tale about reacting to change by attacking the tool.

In Paris around the July Revolution of 1830, printing workers destroyed mechanical presses amid upheaval and concerns about livelihoods. In York, Upper Canada, in 1826, members of the political elite smashed William Lyon Mackenzie’s printing equipment and threw type into the water.

Those were different conflicts. The Paris attacks involved mechanization and labour; the York attack was retaliation against a critical newspaper. We shouldn’t pretend there was one universal movement against the printing press.

But both incidents show how machines can become the physical targets of much deeper anxieties about power, employment and control.

Mackenzie sued and won damages. His newspaper survived.

Destroying a press could silence a particular voice or disrupt a particular business. It couldn’t erase printing from history.

That brings to mind one of science fiction’s most ominous lines:

Resistance is futile.

Star Trek fans will recognize the Borg. But I wouldn’t take their advice literally. Resistance to unsafe technology can be necessary and productive. What may be futile is imagining we can solve the challenges of new technology simply by forbidding it to exist.

Commander Data, please take the bridge

Star Trek gives us another, much more hopeful character: Lieutenant Commander Data.

Data was an android serving aboard the Enterprise. Brilliant, dependable and capable, yet his non-human nature repeatedly raised questions about trust and rights.

In the 1989 episode The Measure of a Man, Starfleet must confront whether Data can be treated as property and dismantled against his wishes.

Now imagine being a passenger aboard a starship when a crisis strikes. A highly capable android offers to take command.

Would you feel safer knowing that the person at the controls was biologically human? Or would you care more about competence, judgment, reliability and accountability?

Data is fiction, and present-day AI isn’t established to possess his imagined sentience. But the story asks a useful question: are we evaluating capability and conduct, or are we making assumptions based on origin?

The question may soon extend well beyond websites.

Imagine a robotic guide dog helping someone navigate a shopping centre. A mobile AI assistant accompanying an older adult to an appointment. A humanoid robot carrying bags through an airport. A companion robot attending an event with its owner.

Will the sign on the door say, “You’re welcome, but your robot isn’t”?

There may be valid safety and operational reasons to restrict certain machines in certain places. But if a robot provides an essential service to a person, those decisions deserve far more thought than a blanket prohibition.

What about the real dangers?

There is a serious argument against uncritical enthusiasm for AI.

People worry about job displacement, deepfakes, surveillance, privacy, discrimination in automated decisions, concentrated corporate power, cybercrime and autonomous weapons. They worry about governments and other bad actors using AI to manipulate or control people. They worry about what happens when powerful systems behave in ways their operators did not intend.

Those are not ridiculous fears. They deserve rigorous answers.

Researchers such as Timnit Gebru and Yoshua Bengio have helped bring different questions of fairness, power and safety into public discussion.

The U.S. National Institute of Standards and Technology has also developed an AI Risk Management Framework to help organizations evaluate and manage AI risks.

This work matters precisely because the future shouldn’t be a choice between reckless adoption and total rejection.

The aim is trustworthy technology, meaningful accountability and safeguards proportionate to the actual risk.

I work with the trades more than any other sector, and more of my clients are starting to hand routine online tasks to AI assistants: ordering supplies, checking permit portals, booking appointments. When a website slams the door on that assistant, it’s the contractor who loses the time.

Maybe it’s time for a different question

As technology becomes more intimately connected to us—through phones, watches, glasses, assistive devices and perhaps eventually implants—the boundary between human ability and technological assistance will become increasingly difficult to draw.

At what point does the question “human or machine?” stop being useful?

I’m not suggesting that an AI assistant should automatically have the same legal rights as a person. I’m certainly not suggesting that every automated visitor should have unrestricted access to every website.

I’m suggesting something simpler.

Stop judging a tool solely by the fact that it isn’t human. And stop overlooking the human being who may be depending on it.

Protect websites from fraud. Protect students from poor educational practices. Protect society from unsafe systems and people who misuse them.

But make room for authorized agents, accessibility, thoughtful innovation and collaboration.

The future I would rather see isn’t one where machines replace humanity. It’s one where technology helps more people participate, create and contribute—sometimes in ways that would otherwise be impossible.

So the next time a website asks me to tick the box saying “I’m not a robot,” I’ll probably tick it.

But I’ll also be thinking about the AI assistant I might have sent in my place.

And I’ll wonder whether the better question was never “Are you a robot?”

Perhaps it should have been:

“What are you here to do—and whom are you here to help?”