Is There Such a Thing as Ethical AI?

In a cafe tucked under a mall escalator, Carljoe Javier sets his coffee down to sweep his moppy hair off his face, revealing a sharp undercut. In a dark t-shirt, he suddenly looks like a guy who plays guitar in a rock group–and it turns out he is. He’s heading to band practice right after this. But first, as Chief AI Ethicist of the non-profit Data and AI Ethics PH–whose core tenet is that "All AI should be responsible”–he’s going to answer a few questions I have about artificial intelligence, its uses, misuses, and abuses.

Then, just as quickly as my notions of what hair and hobbies an AI Ethicist should have are thwarted, the interview I expect to conduct gets derailed. At the top of my notes is an accusation, more than a question: How are you okay with being a proponent of AI when it’s destroying the planet? He chuckles graciously and says the thing that makes me close my notebook of accusations and listen with genuine curiosity. "AI problems are human problems."

He would know. An award-winning speculative fiction writer with a lifelong interest in philosophy, Carljoe–also called Carl–brought his two interests together during a career break, and got certifications in Responsible AI from the University of Edinburgh and Wharton. This opened the door to his current role, where, among other things, he holds discussions for government agencies, international organizations, academic institutions, and the general public advocating for not just efficient but ethical uses of AI, particularly in the Philippine context.

Still, expert or not, the question remains: In an era where the most revolutionary technology in recent history is burning through our climate, critical thinking skills, and career options, is there even such a thing as "ethical AI"?

ALSO READ:

AI Is Proof of How Incredibly Lonely We Really Are

Coping With the Impending AI Talent Takeover (With Pure Delusion!)

The Case Against AI

Let’s be honest: the criticisms against artificial intelligence are valid.

The environmental toll of building and running the massive data centers that process AI is staggering, especially as Silicon Valley companies plant hot, power-hungry machines in desert regions that require huge amounts of water to function.

At the same time, there is widespread worry about how AI is disrupting the labor market. On the creative end, artists and writers raise alarms over generative AI models trained on copyrighted content without consent, compensation, or credit. While on the corporate end, large swathes of employees are losing their livelihoods as companies adopt AI-first business models.

There’s also the issue of digital colonization, which Carl himself is preoccupied with. As most AI tools are trained, owned, and hosted in the United States or China, Filipino (and other nations’) data is processed and stored by these superpowers, with no local oversight. This lack of AI sovereignty leaves us at the mercy of their self-interest.

Then there’s the fear of critical thinking eroding–students relying on AI to do the tasks that traditionally develop one’s ability to reason, professionals copy-pasting ChatGPT outputs without understanding or verifying what they're delivering, governments executing AI-driven policies that are admittedly flawed.

For the more technologically minded, another danger is baked into AI’s very foundation: data. Researchers found that because the datasets AI is trained on are often full of racial, gender, and cultural biases, AI risks becoming another tool of systemic oppression–unless the datasets are deliberately examined and dismantled.

The Importance of Asking

With all these problems, isn’t it better to put the genie back in the bottle and rid ourselves of AI for good? Carljoe responds to the question with one of his own: "I respect people who say they will not touch this tech for all these reasons–but is it the technology itself, or is it the way we built it?" It’s less a knee-jerk pilosopo reaction and more a sincerely philosophical query.

Expanding on the idea that AI problems are human problems, Carl adds, "We’ve been building AI through the Silicon Valley model, which is revenue-driven. But we don’t need to continue that way." The problem, therefore, isn’t AI itself, but "the capitalistic system that is willing to forego humanity and the environment to improve profit margins."

Regarding the environment, Carl cites issues like the US’s 10-year moratorium on AI regulation as an accelerant of the planet’s degradation–not the tech itself. Regarding labor, especially artistic labor, he observes how society’s long-standing devaluation of workers and of art, are the root of the problem–not the tech itself. Regarding technocolonization, he points out that it’s the superpowers who are in an "arms race" to consolidate power–not the tech itself. Regarding critical thinking, he notes how the traditional educational system is what drives students to center their learning around compliance and employability rather than analysis and reflection–not the tech itself. And regarding biased datasets, he nods. It’s the data curation, perhaps the curator, that is biased–but not the tech itself.

So, Is There Any Hope?

The situation is not as bleak as it seems, Carl says. His stance is that ultimately, "AI solutions are human solutions." In other words: to fix the problems we ascribe to AI, we will first need to take a hard look at ourselves and our society–and decide to make ethical changes. The dominoes will fall, and AI will follow suit.

In the Philippines, what that could look like is leveraging AI’s unmatched ability to process large data to efficiently execute our government’s service delivery systems, thereby improving our institutions, resulting in a more just society. At least that’s Carl’s lofty vision.

One example he gives is with prison reform. In the Philippines, where outdated paper records and decentralized systems trap persons deprived of liberty (PDLs) in jails for years longer than their maximum sentence, AI could help digitize and analyze backlogged data. With proper safeguards, this kind of system could flag where the right to a speedy trial has been withheld and offer the possibility of justice at scale. "AI can crunch that data," Carl says. "It can help free people." And yet, we can’t achieve this without fixing the human problem first: we as a society need to see PDLs as deserving basic human rights and dignity–worthy of reforming our system for.

So, Is AI Ethical?

With our coffee cups empty, Carl heads to band practice, and I haven’t found a clear answer to the question "Is there such a thing as ethical AI?" Mainly, because it’s the wrong question. Or at least an incomplete one.

Asking whether AI is ethical seems akin to asking whether a laptop is ethical. Boiled down to its essence, it’s a technology–a tool. It can be used to empower or exploit, to liberate or oppress. Its potential for good or harm depends on who’s wielding it, and why.

The important thing I’m coming away with is to ask questions at all. Not from a place of accusation, but genuine inquiry and reflection: Should we passively adopt whatever tools Silicon Valley hands us, or should we shape our own standards? Should we reject AI wholesale for its flaws, or should we accept the collateral damage of a world that includes AI?

Ethics and the open-ended, seemingly frivolous, perhaps even irritating questions it comes with, Carl emphasizes, will have more of an impact on the future of AI than straightforward regulation. Navigating moral terrain–how we care for ourselves, each other, the planet, the future–will impact our material reality.

The question “Is there such a thing as ethical AI?” becomes obsolete in the shadow of another, more important one: How do we live ethically?