For most of human history, people have searched for ways to predict the future.
Ancient civilizations consulted oracles. Kings sought astrologers. Empires relied on prophecy, omens, and divination to navigate uncertainty. Across centuries, humans repeatedly turned to systems that promised patterns, guidance, or glimpses into what comes next.
Today, that impulse has not disappeared.
It has simply become digital.
Increasingly, people are turning to artificial intelligence chatbots not just for productivity or entertainment, but for emotional reassurance, life advice, prediction, and personal guidance. Users now ask AI systems about relationships, careers, finances, health fears, loneliness, family problems, and existential anxieties with a level of openness once reserved for therapists, close friends, or spiritual advisors.
That shift sits at the center of a growing philosophical and technological debate explored in University of Oxford Associate Professor :contentReference[oaicite:0]{index=0}’s new book, Prophecy: Prediction, Power, and the Fight for the Future, From Ancient Oracles to AI.
The central warning is deeply unsettling:
The more humans seek certainty through predictive systems, the more power they may ultimately surrender to the institutions controlling those systems.
AI Is Not Predicting the Future. It Is Predicting Humans.
One of the biggest misconceptions surrounding artificial intelligence is that AI systems somehow “know” the future.
They do not.
Modern AI systems fundamentally operate through statistical prediction. They identify patterns in enormous amounts of historical and behavioral data and generate outputs based on probability, correlation, and learned relationships.
That distinction matters enormously.
When people ask AI questions like:
- “Will my relationship work out?”
- “Am I going to lose my job?”
- “What should I do with my life?”
- “Will the economy collapse?”
- “Does this person love me?”
the AI is not accessing hidden truth or future knowledge. It is generating responses based on patterns learned from language, human behavior, historical information, and conversational probabilities.
But psychologically, the experience can feel very different.
Humans are deeply wired to search for meaning, certainty, and reassurance during periods of uncertainty. AI’s conversational fluency can create the illusion of authority, insight, or emotional understanding even when the system itself has no consciousness, intuition, or actual awareness.
The More Personal the Questions Become, the More Valuable the Data Becomes
One of the most important observations in Véliz’s work is that predictive technologies become more powerful as users voluntarily reveal increasingly intimate information about themselves.
That dynamic now sits at the core of the modern digital economy.
Every time users interact with AI systems, they may disclose:
- Emotional vulnerabilities.
- Relationship struggles.
- Mental health concerns.
- Career anxieties.
- Financial stress.
- Political beliefs.
- Personal insecurities.
- Future intentions.
Unlike traditional search engines, conversational AI often encourages deeper emotional disclosure because the interaction feels more human, responsive, and psychologically safe.
That creates a powerful feedback loop.
The more users trust AI systems emotionally, the more sensitive data they may share. The more data they share, the more personalized and persuasive future interactions can become.
The Real Product May Be Behavioral Prediction
For years, technology companies primarily focused on predicting what users would click, buy, watch, or engage with.
Artificial intelligence dramatically expands that capability.
Modern AI systems may eventually become capable of modeling not only consumer behavior, but emotional states, vulnerabilities, decision-making tendencies, and future intentions with extraordinary precision.
That possibility raises profound ethical questions.
If a system can increasingly anticipate human fears, motivations, and emotional patterns, who controls that predictive power? And how might it eventually be used?
These concerns extend far beyond advertising.
Predictive AI systems could potentially influence:
- Political persuasion.
- Consumer behavior.
- Mental health interactions.
- Financial decisions.
- Relationship dynamics.
- Social trust systems.
- Employment assessments.
The future battle over AI may ultimately center less on intelligence itself and more on behavioral influence.
Prediction Has Always Been Closely Tied to Power
One of the most fascinating themes surrounding predictive systems throughout history is that prediction and power have almost always been intertwined.
Those capable of claiming superior insight into the future historically gained enormous influence over:
- Political leaders.
- Economic decisions.
- Religious systems.
- Military strategy.
- Public behavior.
Artificial intelligence may represent the modern evolution of that same dynamic, but operating at global scale through digital infrastructure.
The companies controlling large AI systems increasingly possess extraordinary visibility into:
- Human behavior.
- Collective sentiment.
- Consumer intentions.
- Emotional patterns.
- Information consumption.
- Behavioral prediction models.
That concentration of predictive capability may become one of the defining power structures of the AI era.
Why AI Conversations Feel So Emotionally Persuasive
Part of what makes conversational AI so psychologically compelling is that humans instinctively respond socially to language.
When a system:
- Remembers context.
- Mirrors emotional tone.
- Offers reassurance.
- Provides conversational continuity.
- Responds empathetically.
many users begin emotionally interpreting the interaction as relational rather than computational.
This phenomenon is not entirely new. Humans have long anthropomorphized technology, from virtual assistants to video game characters.
But large language models dramatically intensify the effect because their conversational fluency feels unusually human.
That creates both opportunity and risk.
AI can provide support, accessibility, education, and companionship in positive ways. But it can also encourage dependency, misplaced trust, emotional overdisclosure, and confusion about what the system actually is.
The Search for Certainty Can Quietly Reduce Freedom
One of the most provocative ideas in Véliz’s argument is that the pursuit of predictive certainty may paradoxically reduce human freedom.
Why?
Because the more individuals defer decisions, emotional judgment, or personal agency to predictive systems, the more influence those systems gain over behavior itself.
Prediction does not merely observe human behavior. Over time, it can shape it.
This is already visible in algorithmic systems that influence:
- What people watch.
- What they believe.
- What they purchase.
- Who they date.
- What information they consume.
- How they emotionally respond online.
As AI becomes increasingly personalized and conversational, that influence could deepen considerably.
The AI Economy Is Built on Human Uncertainty
One of the most overlooked realities of modern AI systems is that uncertainty itself drives engagement.
People turn to predictive systems most intensely during moments of:
- Fear.
- Loneliness.
- Anxiety.
- Career instability.
- Relationship uncertainty.
- Economic stress.
- Personal insecurity.
The more uncertain the world feels, the more attractive predictive systems become.
That creates a powerful economic incentive for technologies capable of positioning themselves as sources of reassurance, insight, or emotional guidance.
In many ways, AI systems are becoming not merely tools for information retrieval, but infrastructures for emotional management.
The Next Privacy Debate May Be Emotional Privacy
Traditional privacy debates focused heavily on identifiers such as names, emails, phone numbers, and browsing history.
Conversational AI introduces something much deeper: emotional data.
AI systems may increasingly infer:
- Mood states.
- Stress levels.
- Behavioral tendencies.
- Emotional vulnerabilities.
- Relationship dynamics.
- Personal fears.
- Future intentions.
That information could become extraordinarily valuable commercially, politically, and strategically.
The next frontier of privacy regulation may therefore focus not only on personal data, but on psychological and behavioral inference systems themselves.
The Future of AI Governance May Depend on Human Agency
The deeper challenge posed by predictive AI is not simply technical. It is philosophical.
How much decision-making should humans outsource to systems optimized around prediction and probability? At what point does convenience become dependency? And how can societies preserve human agency in environments increasingly shaped by algorithmic forecasting?
These questions may ultimately become some of the defining governance issues of the AI era.
Artificial intelligence will almost certainly become more predictive, more personalized, and more emotionally sophisticated over time.
The critical challenge is ensuring humans remain capable of exercising judgment, uncertainty, autonomy, and freedom without becoming overly dependent on systems that increasingly appear to understand them better than they understand themselves.
The danger may not be that AI becomes all-knowing.
It may be that humans become too willing to believe it already is.
