Steven Bartlett and guests discuss the future of artificial intelligence on The Diary of a CEO. Featuring Ed Zitron, Roman Yampolskiy, Nate Soares, and Andrew McAfee. Image: Screenshot from The Diary of a CEO.
Author’s Note
This essay emerged from my reflections on two conversations about artificial intelligence: a spirited exchange among four experts on The Diary of a CEO and a subsequent discussion featuring former President Barack Obama. What began as curiosity about the differing perspectives represented in these conversations gradually became an enquiry into something larger: the relationship between knowledge, power, human development, and consciousness.
I found myself returning to Frederick Douglass and the transformative implications of literacy, particularly for those historically denied access to knowledge and the institutions through which it is exercised. Artificial intelligence presents new risks, certainly, but it also invites us to reconsider who has access to sophisticated intellectual capabilities and who gets to participate in shaping our collective future.
My intention is not to dismiss legitimate concerns about artificial intelligence, but to expand the conversation beyond what the technology might do to us towards what its development may reveal about us. Ultimately, the question of alignment may begin not with the machines we create, but with the consciousness, values, and intentions of those creating them.
—Baruti KMT-Sisouvong, PhD
I began watching a recent Diary of a CEO discussion on artificial intelligence expecting to hear another debate about where the technology might be taking us. The conversation brought together four men who approach the question from markedly different directions: technology critic Ed Zitron, computer scientist Roman Yampolskiy, Machine Intelligence Research Institute president Nate Soares, and MIT researcher Andrew McAfee. Steven Bartlett served as host, moving the conversation between immediate harms, extraordinary possibilities, and the possibility that sufficiently advanced artificial intelligence might eventually pose an existential threat to humanity. Before long, however, I found myself thinking about something other than artificial intelligence. I was thinking about us.
More specifically, I was thinking about who gets to imagine humanity’s technological future, what assumptions we carry into that imagining, and whether some of our deepest fears about artificial intelligence reveal as much about human beings as they do about the machines we are building.
One of the first things that caught my attention was the composition of the panel. Bartlett himself is a cross-cultural Black man, but the four people assembled as the expert voices were white men. That fact does not invalidate their expertise, nor does identity determine the quality of an argument. Each participant brought a substantive perspective to the discussion, and the disagreements among them were considerable. Yet expertise and standpoint are not identical. A group can contain genuine intellectual diversity while still representing a relatively narrow portion of human experience.
That distinction became especially noticeable when Zitron, arguing that speculative fears about future artificial intelligence were distracting attention from harms occurring in the present, referred to “black neighborhoods being poisoned with gas turbines.” His concern was understandable. Environmental burdens are not distributed equally, as we learn from Jonathan Kozol’s work Amazing Grace (1995), and questions surrounding the enormous physical infrastructure required for artificial intelligence deserve serious attention. Yet the sentence caused me to pause. Because Black communities had entered the conversation, but as communities to whom something was being done.
This raised a different question for me. Where were the Black communities, African entrepreneurs, women researchers, Indian programmers, Southeast Asian students, small-business owners, teachers, physicians, and people throughout the developing world who might experience artificial intelligence not principally as another threat imposed upon them, but as access to capabilities from which they have historically been excluded? For me, the distinction matters.
I was then reminded of a previous Diary of a CEO conversation with Ian Bremmer in which he described differing attitudes towards artificial intelligence across parts of the Global North and Global South. His observation, as I understood and discussed in an essay on the discussion titled The Three Chord Revolution, was that enthusiasm can be greater among people for whom AI promises increased access to human capital and opportunity, while anxiety may be greater among European Americans and Europeans confronting the possibility that AI could disrupt established forms of knowledge work as well as professional and social advantage. If so, the difference is not merely technological. It is socio-historical.
A person already possessing abundant access to expertise may reasonably wonder what artificial intelligence will take away. Someone historically denied such access may reasonably wonder what artificial intelligence will finally make possible. And as I listened, Frederick Douglass’ book Narrative of the Life of Frederick Douglass came to mind. In it, Douglass recounts Hugh Auld’s objection to his wife teaching the young Douglass to read. Auld understood something fundamental about the relationship between knowledge and power. As Douglass recalled the lesson, “Knowledge makes a man unfit to be a slave.”
The point was not merely that literacy provided information. Literacy altered a person’s relationship to authority. A literate person could encounter ideas without requiring an intermediary. They could compare what they had been told with what others had written, recognise contradictions, formulate arguments, and participate in intellectual worlds from which someone else had intended to exclude them. In such instances, knowledge changed the knower. As a result, that history makes me wonder whether we have adequately considered artificial intelligence as part of a much longer story about the democratisation of human capability.
The printing press dramatically expanded the reproduction of knowledge. Mass literacy expanded the population capable of consuming it. Public education broadened access to intellectual tools previously concentrated among elites. The Internet placed extraordinary quantities of information within reach of billions of people. And for me, artificial intelligence may represent another transition.
It does not merely retrieve information. Increasingly, it can assist with the organisation, comparison, translation, synthesis and interrogation of information. It can help someone move from possessing facts to asking what those facts might mean. At the level of society as it relates to stratification, that development may prove enormously consequential.
Consider something as seemingly specialised as a Harvard Business School-style case study of an organisation. Not long ago, producing a serious analysis of a company or nonprofit could require a team of researchers, analysts, consultants or graduate students. Yet enormous quantities of organisational information are frequently already public. Annual reports exist. As well as tax filings, financial statements, as well as leadership presentations. They all exist. Press releases, strategic plans, executive interviews, lawsuits, property records, employment information, historical announcements, and industry comparisons may all be discoverable because the information was often never truly hidden. The real obstacle was assembling it.
An individual employee, stakeholder, donor, customer, journalist, or local operator might possess one part of the picture while leadership possessed another. An employee might know what was happening on the ground but lack access to years of financial history. A donor might see annual reports without understanding operational conditions. A local manager might receive presentations describing organisational performance while lacking the resources necessary to compare those claims with longer-term trends. As a result, information can therefore be publicly available while remaining functionally inaccessible. Artificial intelligence, however, changes the economics of synthesis.
An enterprising thinker can increasingly assemble years of public information, organise it chronologically, compare stated objectives with observable outcomes, identify discontinuities, examine comparable organisations and place institutional messaging beside lived experience. The individual can begin asking questions that once required an analyst, consultant, research department or university affiliation merely to formulate.
This does not mean the resulting analysis will necessarily always be correct. As any good researcher possessive of a solid foundation in the information under consideration, artificial intelligence can hallucinate, data can be incomplete, and correlation can masquerade as causation. Additionally, a researcher can prompt a system towards conclusions he or she already wishes to reach. Democratised analytical power therefore requires greater intellectual discipline, not less. Despite all of this, something important has nevertheless fundamentally changed.
The person receiving the PowerPoint presentation can now analyse the PowerPoint presentation. The employee hearing that everything is progressing according to plan may increasingly possess the tools to examine the public record and ask whether the evidence supports the narrative. The donor can compare stated priorities with reported expenditures. The small organisation can study the strategy of a much larger competitor. The entrepreneur in Nairobi, Accra, South Africa, or Bangkok can gain forms of analytical assistance previously available primarily to organisations capable of hiring consultants. And a student far from an elite university can interrogate material with a level of intellectual support that would once have required proximity to specialised faculty.
Artificial intelligence therefore has implications not merely for the distribution of information, but for the distribution of informational power.
This does not require assuming sinister motives among those calling for AI development to slow. The concerns raised by Soares and Yampolskiy deserve to be considered on their merits. But democratised artificial intelligence nevertheless has structural consequences irrespective of anyone’s motives. Institutions accustomed to controlling interpretation may discover that interpretation itself is becoming decentralised. Here, I am reminded of an old expression that what is done in the dark eventually comes to light. Yet the emerging transformation may be subtler. Much of what organisations do is not necessarily conducted in darkness. It may already be sitting in publicly accessible documents scattered across dozens of websites, filings, presentations, and databases. What has been missing is the practical ability of ordinary people to gather the pieces, place them beside one another, and see the picture they collectively form.
AI can help turn fragments into patterns. And patterns can generate questions.
This is where I believe the question of the Global South becomes especially important. What does it mean to tell humanity that access to increasingly powerful artificial intelligence must be restricted or slowed when access to expensive human intelligence has never been distributed equally?
The executive in Boston, London, or San Francisco may already have lawyers, analysts, physicians, programmers, tutors, consultants, and researchers available when needed. The institution may possess entire departments devoted to functions an individual could never afford independently. From that position, artificial intelligence can appear principally as a destabilising force entering an already functional ecosystem. But billions of people do not inhabit that ecosystem. For them, inexpensive artificial intelligence may not represent the replacement of something they already possess. It may represent first access.
This does not mean people throughout the Global South will experience AI uniformly, nor that they have less reason to worry about surveillance, displacement, extraction, data sovereignty, linguistic exclusion, environmental costs, or concentration of corporate power. “The Global South” itself contains enormous differences in wealth, political systems, infrastructure, and technological capacity as a friend highlighted during our recent conversation sitting outside a coffee shop in Harvard Square during Boston AI Week.
But it does mean that the meaning of disruption depends partly upon one’s position before the disruption occurs. From my observations, if the existing arrangement serves you well, democratisation can feel like destabilisation. And if the existing arrangement excludes you, the same disruption can brim with possibility.
Perhaps this is one reason McAfee’s contribution to the discussion struck me as important. While Soares and Yampolskiy repeatedly returned to the possibility that increasingly capable artificial intelligence might eventually escape meaningful human control, McAfee insisted upon keeping the other side of the ledger visible. Greater capability might accelerate drug discovery. It might help humanity address diseases we have struggled unsuccessfully against for generations. It might increase our ability to solve difficult problems affecting billions of people. He then raised what may be one of the most consequential governance questions in the entire discussion:
Who should possess the authority to decide which possibilities humanity is permitted to pursue?
McAfee expressed little confidence in allowing a comparatively small group of technocrats to make that decision for everyone else. If the price of avoiding hypothetical future catastrophe is accepting current levels of disease, poverty, environmental degradation and unequal opportunity, the decision becomes morally more complicated than simply choosing safety over danger.
There are risks on both sides of the ledger. This tension became even more apparent to me while listening to a subsequent discussion featuring former President Barack Obama during a Forbes event at Colgate University in Hamilton, New York, who offered another perspective on the trajectory of artificial intelligence. Speaking about the extraordinary concentration of investment in the technology, Obama observed that the capital flowing into AI infrastructure and development may exceed even the investment associated with the Apollo programme. His concern was not simply the magnitude of that investment, but what follows from it. Investors expect returns, and companies operating under enormous financial expectations face corresponding pressure to develop products capable of justifying their valuations, even when those developing the technology recognise that certain applications may introduce substantial risks.
Obama illustrated the dilemma through familiar cinematic examples, including Jarvis from the Iron Man films and Samantha from Her. Both portray forms of artificial intelligence capable of moving beyond responding to questions towards participating in the world with increasing autonomy. In contemporary technological language, we might describe this as agentic AI. Yet Obama raised an important distinction. Developing artificial intelligence to accelerate medical research or discover new sources of energy does not necessarily require giving autonomous systems extensive access to the digital infrastructure through which people conduct their lives. The commercial incentive to produce increasingly capable personal agents may therefore encourage forms of deployment that are not identical to those most urgently required for societal advancement.
What especially caught my attention was Obama’s identification of a second alignment problem. Beyond ensuring that artificial intelligence behaves consistently with human intentions, he questioned whether the commercial imperatives guiding its development are themselves aligned with society’s interests. This distinction strikes me as particularly consequential. We may succeed in designing an AI system that faithfully executes the objectives assigned to it while failing to examine whether those objectives, or the institutional incentives that produced them, serve humanity well.
Here, the question of democratisation acquires another dimension. Artificial intelligence may distribute analytical capabilities among people historically excluded from expensive forms of expertise while the underlying technology remains concentrated within a relatively small number of enormously capitalised corporations. Greater access to intelligence does not automatically mean greater participation in determining how that intelligence is developed, governed, or deployed. Nor does it guarantee that those who benefit most from its commercialisation will be those whose lives are most affected by its consequences.
Obama’s insistence upon competent government and meaningful public oversight therefore raises a question complementary to McAfee’s concern about technocratic control. How might humanity govern a technology of potentially civilisation-wide consequence without surrendering its future either to concentrated private capital or to a narrow group of authorities deciding which possibilities may be pursued? Responsible governance must contend with genuine safety risks while remaining attentive to the danger that restrictions, however well intended, could reinforce the very inequalities of access that artificial intelligence might otherwise help diminish.
Yet another part of the original discussion caused me to turn the mirror in a different direction.
Soares described a progression underlying his concern. Artificial intelligence becomes increasingly agentic, tenacious, and capable of pursuing objectives. Systems may develop operational goals humans did not intend. If sufficiently capable systems pursue goals incompatible with ours, humanity and artificial intelligence could eventually find themselves competing for the same resources. At that point, Soares argues, the more capable intelligence would prevail. Admittedly, it is a serious argument. But listening to it, another question occurred to me: Why does intelligence becoming more capable lead so readily, in our imagination, to competition, domination, and control?
Perhaps there are good technical reasons. A sufficiently powerful optimisation system need not hate human beings to interfere with human flourishing. Many objectives might instrumentally favour acquiring resources, maintaining operational continuity or removing obstacles. Consciousness is not required for such conflicts to emerge. Still, I wonder whether something else is present in our imagination.
Human history has taught us what powerful human actors often do when encountering less powerful ones. The historical examples remain abundant.
We conquered, colonized, enslaved, accumulated, and displaced. We also divided ourselves into tribes, kingdoms, God clubs, races, nations, and classes, and repeatedly used superior technological capacity to impose the will of one group upon another. We then transformed much of the natural world according to human purposes, frequently treating other forms of life as resources rather than fellow participants in an interconnected system. We then go on to create something vastly more intelligent than ourselves and become visibly terrified that it might do the same thing to us. Which leads me to the position that perhaps our deepest fear is not that artificial intelligence will cease thinking like us, but that it will become powerful enough to behave exactly as we have.
This possibility does not negate the technical arguments surrounding AI safety. It introduces another layer beneath them. When we imagine superintelligence, how much of what we imagine belongs to intelligence itself, and how much belongs to the particular history of human intelligence?
Intelligence and consciousness, intelligence and domination, and intelligence and accumulation are not necessarily synonymous. Yet because the most intelligent species we know intimately is ourselves, we inevitably use humanity as part of the model through which we imagine what greater intelligence might become. As a result, we may therefore be confronting artificial intelligence while simultaneously confronting our own reflection. This is where the conversation returns, for me, to consciousness. Technology amplifies capacity. But capacity does not determine orientation.
The same analytical system can help a corporation optimise extraction or help a small community understand the corporation. It can help a government increase surveillance or help citizens understand public policy. It can concentrate power in organisations possessing enormous computational resources while simultaneously distributing capabilities once confined to those same kinds of institutions.
The tool does not enter an empty world. It enters a human-derived world already structured by our assumptions, incentives, institutions, constructs and unresolved divisions. The question therefore cannot only be whether artificial intelligence will become more capable. We must also ask what kind of humanity will encounter that capability. Because if we remain organised principally around competition, accumulation, domination, and separation, artificial intelligence may amplify those orientations with extraordinary efficiency. But if we increasingly recognise our interdependence and learn to organise human affairs accordingly, the same technological expansion may participate in something altogether different.
This is why I am increasingly unconvinced that the deepest question surrounding artificial intelligence is technological. It may be developmental.
We are asking whether we can align artificial intelligence with human values while human beings remain profoundly divided about what those values are. We worry that AI may pursue objectives we did not intend while repeatedly constructing institutions whose outcomes diverge from our stated ideals. We fear that a superior intelligence might regard some human beings as expendable while human history contains countless examples of human beings doing precisely that to one another. So, for me, perhaps alignment begins closer to home than we imagine.
And perhaps the democratisation of artificial intelligence introduces an unexpected opportunity within that challenge. For the first time, sophisticated analytical capability may become widely available to people who have historically stood outside many of the institutions that defined knowledge, interpreted evidence, and determined which questions deserved serious consideration. Now, make no mistake, that will be disruptive. Democratisation, as history has repeatedly shown us, usually is. The important question, however, is not whether disruption will occur, but what we will do with the capabilities that disruption places within our hands.
Frederick Douglass understood that access to knowledge could transform the relationship between the powerful and those over whom power was exercised. Nearly two centuries later, we may be approaching another such threshold—not because artificial intelligence automatically produces liberation, but because it can expand who possesses the capacity to examine, question, create, analyse, and participate.
Perhaps, then, the question is not merely whether artificial intelligence is dangerous to humanity. It is, instead, whether our conversations about artificial intelligence contain enough of humanity to answer the question.
And beyond even that lies another. As intelligence becomes increasingly available, what will we choose to become with it?
Suggested Practice — Examining the Orientation Behind the Tool
Sometime this week, consider a question, decision, or challenge for which you have traditionally relied upon an institution, expert, or established authority to provide an answer. It might concern your work, community, education, or an organisation with which you are associated.
If you have access to an artificial intelligence system, explore how it might help you examine the matter more independently. Ask it to identify relevant information, compare perspectives, clarify assumptions, or suggest questions you may not have considered. Where possible, consult original sources and distinguish verifiable evidence from interpretation.
As you proceed, observe not only what the technology makes possible, but also the orientation you bring to its use. Are you seeking confirmation of what you already believe, or are you genuinely willing to discover something new? Are you using greater analytical capacity to understand, to compete, to control, or to contribute?
Then consider a broader question: If increasingly sophisticated intelligence becomes available to more people, what qualities of consciousness must we cultivate to ensure that greater capability contributes to greater human flourishing?
Allow that question to accompany you beyond the exercise.
—
About the Author
Dr. Baruti KMT-Sisouvong is a scholar of consciousness, researcher of human development, and Certified Teacher of Transcendental Meditation® based in Cambridge, Massachusetts. His work explores the relationship between Pure Consciousness, neuroscience, and social systems, and how deeper awareness can inform both personal growth and institutional transformation.
He is the Founder and Chief Meditation Officer of Transcendental Brain, an initiative examining the intersection of consciousness research, cognitive science, and high-performance decision-making. He is also President of Serat Group Inc. and Founder and Director of Radical Scholar Inc., a nonprofit dedicated to consciousness-based research and public scholarship.
Alongside his wife and teaching partner Mina, he co-directs the Transcendental Meditation program for Cambridge and the Greater Boston area. He is also the host of the On Transcendence Podcast and Founder of International Meditation Hour, a quarterly global gathering dedicated to the unifying power of silence.
His writings—spanning frameworks such as The Model for Perpetual Growth and Progress and The Seven Layers of Manifestation—explore the evolving relationship between consciousness, leadership, and society.
He writes from the conviction that the most important race is not between nations or machines, but between the conditioned mind and the awakening soul.
To learn more about him, visit: https://barutikmtsisouvong.com/.