The Advent of AI: Three Propositions

On Saturday, I enjoyed the honor of joining new Mozilla Foundation Executive Director Nabiha Syed and an amazing group of Mozilla leaders to address the first two classes of Mozilla's Rise25 award winners in Dublin. As we rapidly develop and deploy new artificial intelligence technologies, these leaders are at the forefront of efforts to ensure that technology serves rather than imperils efforts to build a more just global society.

In my remarks, I argued for three ideas about the risks and potential rewards of new technology:

1. That technological "invention" is not the same thing as "innovation."
2. That no technology is ever, by itself, morally neutral.
3. That the greatest power of technology is to reveal new possibilities for how our world can be organized.

I've excerpted my full remarks below.

###

Remarks to the Mozilla Rise25 Fellows, Dublin, August 10, 2024

In considering broadly the panoply of challenging technical and ethical questions attending the advent and further development of the loose set of technologies categorized as “artificial intelligence,” our founder Doris Duke is an unlikely and yet highly appropriate figure.

She lived from 1912 to 1993, witnessing the whole of the communications revolution from wireless radio and fixed landline telephony to the personal computer. Two more years and she would have seen the birth of the commercial internet.

More importantly, perhaps, she witnessed the convulsions of early 20th century modernity yield two catastrophic world wars before giving way to a highly imperfect but largely stable global liberal order that had reached its ascendancy around the time of her death.

Today, we are living in what appears to be the implosion of that order—with AI playing an important role in both confounding the institutions that emerged from the wreckage of the second world war and, almost inevitably, in shaping what order is to come.

If it wasn’t already clear to you, these are the stakes facing the Mozilla Rise25 awardees.

It is in you that we are investing our hopes that not only is the next global order more stable, more just, more sustainable—but also that technology like AI that will play a salutary rather than corrosive role.

So what could you, those entrusted with such an exalted and contemporary responsibility, learn from a 20th century scion like Doris Duke?

Someone who likely could not have dreamed of the technology we carry around in our pockets, much less the astounding software and hardware many people in this room are imagining and building?

In her life, her tastes and her philanthropy, Doris Duke dared to be modern and dared to be different.

By modern, I mean in the technical sense. She was always of her time: she funded Planned Parenthood in the 1930s; she supported Historically Black Colleges and Universities in the 1960s. Doris Duke was not nostalgic.

And she was unafraid to buck convention in the service of some beauty, whether moral or artistic. She supported early jazz and contemporary dance artists when these were not mainstream performing arts; she supported research on HIV/AIDs in the 1980s—when many world leaders wouldn’t even concede the disease was real.

In the way she lived her life, Doris Duke taught us something about what innovation is. And that’s what I want to discuss for just a few moments this evening.

In particular, I want to encourage you to consider three propositions that are not specifically about artificial intelligence but that I think get at the heart so many of the vexatious issues that we confront in the early days of AI:

• Proposition One: that innovation and technological development are related but not the same thing.
• Proposition Two: that no innovation—and especially not technology-enabled innovation—is ever morally neutral.
• Proposition Three: that the greatest benefit of technological development to innovation is its ability to reveal the contingency of human circumstance.

Let’s start with proposition one: that innovation and technological development are not the same thing.

We’ve become used to the idea that there is no innovation without technology. That innovation and technological invention are the same thing.

Certainly, this is an idea that many of the corporate operators of artificial intelligence want us to believe because it allows them to argue that any negative constraint, such as regulation, imperils innovation itself.

But I tend to think of innovation in the same way as the 15th and 16th century philosopher Niccolo Machiavelli.

Machiavelli is mostly remembered today as the father of realpolitik, but it’s probably more accurate to think of him as the first philosopher of modernity: of the idea that through reason, knowledge and power we could change our condition.

One of the products of reason and of knowledge that confers a kind of power is technological innovation.

But there’s lots of other kinds of innovation, too—in how we live and associate, in our modes of economic production, in our politics.

And it’s the interplay between these kinds of innovation—all driven by an abiding loyalty to reason—that is the hallmark of modernity.

That’s why Machiavelli talks about innovation as the giving of “new orders”: innovation is a new way to organize social, political and economic life so that we achieve qualitatively different outcomes.

This is the kind of innovation I see in the work of 2023 Rise25 winners like Andy Yen, who is reimagining what the internet looks like if privacy were the default. Or 2024 awardee Raesetje Sefala who is using AI to examine how apartheid’s physical organization still haunts and shapes the South African society of today.

I also see this kind of innovation in the work of 2023 honoree Keoni Mahelona, who is using technology to promote the recovery of indigenous languages and 2024 winner Kathleen Siminyu who is working to ensure African languages are better represented on digital platforms.

These projects are not applying artificial intelligence technology in ways that entrench and reinforce the logic, the habits, the defaults of the world as it is. They are using technology to help us glimpse a different world.

We need innovators because it is innovators who illuminate the gap between the frontiers of our moral imagination and the present capacities of our institutions.

Innovators can also be technological inventors. But technological inventors who do not seek to see beyond and transcend the limits of the present way of organizing our society are not innovators.

Let me put a fine point on this: a mobile ride share platform that increases the intensity of transportation labor is not by itself innovation; a better way to promote consumption by more precisely and subtly manipulating consumer preferences is not by itself innovation; technology that displaces jobs rather than enhancing them is not by itself innovation.

These examples lead right into the second proposition: that no innovation—and especially not technology-enabled innovation—is ever morally neutral.

A sentiment voiced by many technologists and technology observers in commercial and nonprofit sectors is that technology is itself neutral—that morality is only implicated in how technology is used.

Put aside the feeling that in light of the present pathological uses of AI—such as revenge porn or political misinformation—this feels a little bit like “AIs don’t kill people; people kill people.”

Because there’s something not just distasteful but analytically wrong with this worldview.

Since the first industrial revolution, there has never been an intermediate capital good that existed outside of a prevailing political economy.

A less pretentious way to say this is that we build technology inside human societies and we embed human aims and ends in every aspect of the technology we build.

Instagram will be dangerous for the mental health of teen girls because it inhabits an economic system built on asset appreciation and therefore requires intense user engagement—by any means possible—to increase market capitalization. Amazon warehouses risk being uncomfortable and dangerous places to work as long as the human need to excrete is seen as “friction” in the production of gross merchandise value.

What’s important is not just that technology will inevitably be coopted by the prevailing social, economic and political norms of its time. It is that those norms will shape how we go about designing technology itself, including who and what we design technology for.

A particular worldview is inevitably encoded—literally and figuratively—into the defaults that shape digital tools like artificial intelligence.

That’s why we need the work of people like 2023 Rise25 winner Natalia Domagala, who helped conceive of the idea that there should be a public data ethicist or 2024 honoree Deborah Raji, who studies how we can evaluate AI applications for bias—that is, how we can uncover and make visible encoded human fallibilities.

Or 2023 awardee Ahmad Hegab whose work seeks to address technology-enabled gender-based violence and 2024 honoree Elaine Nsoesie who has used technology to help us grasp the structural nature of racism.

These leaders recognize that artificial intelligence will not be conceived outside of our existing cultural and cognitive horizons. That harvesting the positive potential of technology first requires a sober assessment of our existing institutions and the agendas they are designed to serve—and then an ongoing ability to interrogate whether and how technology is entrenching rather than dismantling those structures that we seek to transcend.

Liberatory rhetoric has never alone satisfied the demands of justice. Technology must materially prove its contribution to moral progress. It must withstand the scrutiny of the real world.

This brings me to the third proposition: which is that technology’s greatest contribution to innovation is simply to throw into stark relief the mutability of our world—and therefore its capacity for improvement.

As I’ve sought to argue, technology does not alone represent innovation—but it does represent the modification of human potentialities.

Technology can reduce the human energy cost of many tasks. Technology can reveal layers of the natural world unavailable to our human senses. Technology can, in some cases, speed along or even surpass human cognition.

Technology is all that we add to the natural world. And its secret power is to reveal how little of the world is truly natural.

The elision of subjective human interests and the fixity of nature has long been one of the most powerful ways human beings seek to control one another.

Because to declare something is “natural” or is a “fact”—whether that thing is how human beings love each other or how groups of people with visibly apparent skin tones relate to each other—is to place that thing beyond the reach of social change.

Nature is what is. Social change reaches for what ought to be.

But technology—even when it falls short of its professed aims—can help to unmask the ought parading as an is.

Let me give you an example from Doris Duke’s own life.

In 1966, Duke provided the equivalent of $10 million dollars today to found the first and still largest archive of Native American indigenous oral traditions at a set of universities around the country.

As one of the professors involved in the project recalled, Duke’s “argument was that the library was filled with books about Indians [sic], all written by non-Indians [sic]…She believed that the tape recorder would give these otherwise voiceless people an opportunity to talk.”

In 1963, Phillips had just introduced the Compact Cassette—what people of my generation called a cassette tape.

Duke’s belief was that making audio recording portable would better connect knowledge producing and archiving institutions like universities to tribal communities.

Of course, it didn’t work out exactly as Duke had intended.

Researchers did collect over 6,000 oral histories during the initial life of the project. But the dream of indigenous control over their voices and histories remains elusive—it is researchers, not native communities, who still manage the archive and who enjoy the most access to the collection.

But the point of the story is that technology gave Doris Duke—and us—a new target at which to aim. The impracticality of direct engagement with tribal populations naturalized the appropriation by universities of the exclusive right to tell the authoritative story of Native American communities.

A new technology—the cassette player—illuminated for Duke that this was not natural. It was simply a default choice within the context of existing technology.

And the failure of the project to realize her aspiration in no way undermines the moral warrant for that aspiration.

Technology helped Duke—helped all of us—to see what ought to be.

Indeed, the principal aim of the recent stages of the project—which continues to this day—has been to restore indigenous control over the recordings and oral histories.

And this is why we need the work of people like 2023 awardee Marlena Myles, whose art can enable us to see the world through the eyes of indigenous populations. Or 2024 honoree Stephanie Dinkins, who has produced immersive work that helps us to encounter the resilience of Black women across disparate times and places.

They are showing us a different version of what is, illuminating a rich new landscape of ought.

Taken together, these three ideas—that innovation is different from technology, that technology can never be neutral, but that technology can expand our sense of the possible—are essential ingredients if we want a world in which technologies like AI are a lever for improvement in the human condition, a world in which technology is subordinated to serve the public interest of the many, rather than the partial interest of the few.

And we’re counting you, the Rise25 winners, to show us how to make that world possible. The velocity of AI development, its technical abstruseness, and its concentration in the hands of a few companies threatens our collective agency over the systems, institutional and otherwise, that we erect to fulfill our aims.

You have the aptitude and the ingenuity to interrupt what technologist and critic Judy Estrin has called a “culture of inevitability” regarding  the role of AI in our society.

You can harness technology not for its own sake, but as the fountainhead of new orders—new ways of organizing our society and our world.

Let me close with two thoughts.

The first is a hope.

It can sometimes feel like the techno-deterministic forces that many in this room are challenging are simply too large, too powerful, too overwhelming.

But as I was walking along the River Liffey just this afternoon, I became cliché—recalling the legendary but futile Easter Rising of 1916 right here in Dublin.

The revolutionary gesture was just that—a gesture, almost farcically overmatched by the might of the British state.

But the idea of Irish independence that the uprising birthed was in the end a successful one. And that was because the Easter Rising was what historian Peter Hart once called a “unique example of insurrectionary abstract art.”

It was an uprising that expressed an idea too powerful, too beautiful to be defeated by arms alone. While the rebellion may have failed, as W.B. Yeats immortalized in his classic poem on the uprising: “A terrible beauty is born.”

So the hope is that the work you do becomes a terrible beauty. That you will contribute to an idea about society and technology that will move history, even in the face of seemingly impossible odds.

And the second thought is an admonition.

One weapon we cannot rely upon in the struggle against dystopian visions of the future—whether those visions concern politics or technology—is nostalgia.

We are not here to rescue some original and pure idea of digital technology or of the internet that has somehow been subverted.

I have always regarded Mozilla’s commitment to an open web that serves society as an ought. As an always moving target that represents the continuing expansion of our moral horizons.

We do not build a better future by worshipping an idea of the past.

We do so by learning from the past and then devoting ourselves to a version of the future we believe we can reasonably justify to those who follow us.

Thank you for your vision and your leadership. And congratulations on this high honor.