About us
This is a group for those in the Sacramento area who like to discuss politics and philosophy. We have all had different experiences in life, but we always find that political issues are very relevant to our lives and we want to discuss these issues with others. These days, political discussions are sometimes constructive, but unfortunately they often become hostile when people with differing views are in the same room. It is important for us all to hear views other than our own from time to time, and there has to be a way to have honest, thoughtful discussions on issues important to our lives while keeping the discussion calm and polite. In this discussion group, we try to avoid being overly biased when presenting opinions and we will also work to avoid use of fallacies in making points. We expect that there will be differences of opinion, but as long as we keep a philosophical mindset these discussions should not descend into hostility. Many of us also like to discuss philosophical topics and we will have time for this as well and we will probably find that most philosophical concepts have real-world applications, including political. So this discussion group will hover between these two subjects, and we might even at times focus on other subjects as well.
Upcoming events
51
Who Controls the Future? AI, Prediction, Power and Human Freedom
Barnes & Noble, 280 Palladio Pkwy Bldg 913, Folsom, CA 95630, USA, Folsom, CA, USArtificial intelligence will eventually give humanity something we have never possessed before: an extraordinary ability to predict what will happen.
AI can aggregate and analyze enormous amounts of information—our history, biology, movements, relationships, purchases, communications, behavior and larger social trends. As these systems become more capable, their predictions about individuals, organizations and nations may become increasingly accurate.
This raises our central philosophical question:
When AI can predict what will probably happen, does knowing the future give us the moral right to intervene and change it?
And there is a paradox: If an AI prediction causes us to intervene and the predicted event never occurs, was the prediction wrong—or did it accomplish its purpose?
Consider an AI system that predicts that someone has a very high probability of committing a violent crime. At what level of predictive accuracy, if any, should society intervene? Surveillance might prevent a tragedy, but the person has committed no crime. Should civil liberties ever be restricted because of what someone is predicted to do?
Now reverse the situation. Suppose AI predicts that a child has extraordinary intellectual potential. Should society devote additional resources to that child? If we are uncomfortable using AI to predict criminality, should we be comfortable using it to predict genius?
The same dilemma could arise in health. Suppose AI can predict cancer, addiction or serious mental illness years before symptoms appear—and eventually knows how to prevent or treat it. Does society merely offer that intervention, or can circumstances ever justify requiring it? What happens when an individual's right to refuse treatment conflicts with costs or risks borne by everyone else?
Behind all these questions lies an even deeper philosophical one:
Is Free Will an illusion?
What happens to our concept of free will if human behavior becomes increasingly predictable?
If a sufficiently capable intelligence could predict our decisions from our biology, experiences, environment and behavior, would that suggest that our choices are largely determined? Or would our ability to learn the prediction and deliberately act differently demonstrate that genuine choice remains?
This raises our central political question:
Who Controls the Intelligence?
Even if AI becomes extraordinarily powerful, AI itself may not initially be the greatest political problem. The greater problem may be who controls it.
There are several possibilities.
Governments could control the most powerful AI. This could give states unprecedented capabilities for scientific discovery, national defense and public administration—but also surveillance, population modeling and social control.
Corporations could control it. A relatively small number of companies, executives and investors could possess capabilities once available only to powerful governments. Governments might regulate those companies, but they might also become dependent upon them.
AI could become decentralized. Powerful systems could be widely available to ordinary individuals rather than concentrated in governments or corporations. That democratizes power—but gives the same capabilities to criminals, extremist organizations and other dangerous actors.
Or perhaps nobody will fully control it. Competition among nations and companies may drive AI development faster than governments or international institutions can regulate it.
This leads directly to the political question:
Will advanced AI strengthen democracy or authoritarianism?
Authoritarian governments may benefit from centralized decision-making, extensive surveillance and fewer restrictions on how AI can be deployed.
Democracies may benefit from open scientific inquiry, competing institutions, entrepreneurship, civil liberties and mechanisms for challenging bad decisions.
But there is a third possibility: AI could cause the two systems to become somewhat more alike.
Authoritarian governments could use AI to create extraordinarily effective surveillance states. Democratic governments could gradually adopt some of the same technologies because they are useful for preventing crime, terrorism, disease or other threats. Meanwhile, private companies may accumulate more information about individuals than governments historically possessed.
The central issue, then, is not simply whether artificial intelligence will become extraordinarily powerful.
It is what human beings will do with the power to predict—and who will be permitted to wield that power.
Questions for Discussion- At what level of predictive accuracy, if any, should preventing harm override an individual's civil liberties?
- Should AI predictions be used differently when they identify positive potential—such as genius or health—than when they predict criminality, addiction or mental illness? Who decides what counts as "positive"?
- If human behavior becomes highly predictable, what would that tell us about free will and personal responsibility?
- Who should control the most powerful AI systems: governments, corporations, individuals, international institutions—or no one?
- Does advanced AI ultimately favor democratic societies, authoritarian societies, or something entirely different?
- Most importantly: How do we prevent whoever controls the most powerful AI from using that intelligence primarily to preserve and expand their own power?
Sources:
- Bernard Harcourt — Against Prediction: Profiling, Policing, and Punishing in an Actuarial Age. This may be the most uncannily perfect source for your topic—and it predates today's generative AI. Harcourt asks what happens when government becomes increasingly good at predicting who will commit crimes. His deeper argument is that even accurate prediction can distort justice because we gradually move from judging people for what they have done to managing people according to what statistical models say they are likely to do.
Overview of Against Prediction from Columbia Law School
2. Shoshana Zuboff — “Surveillance Capitalism or Democracy?” This is the one I'd use for your billionaires/corporations versus government question. Zuboff's insight is that companies don't merely collect information about us. Human experience becomes raw material for systems designed to predict behavior, and prediction itself becomes economically valuable. Her argument goes one uncomfortable step further: the commercial incentive eventually moves from predicting behavior toward influencing it. That raises a delicious question for your group: Who has more power—the institution that can punish your behavior or the institution that can quietly shape it?Surveillance Capitalism or Democracy?
3. Daron Acemoglu & Simon Johnson — Power and Progress. This provides a needed counterweight to technological determinism. Their historical argument is essentially: technology doesn't decide who benefits from technology—institutions and political power do. They look across roughly a thousand years and argue that technological advances have repeatedly enriched narrow elites unless societies deliberately created institutions that distributed their benefits. That takes your discussion from "Will AI control us?" to the much more practical question, "Who gets to decide how AI is deployed?"Power and Progress — MIT
4. Katia Schwerzmann — “Abolish! Against the Use of Risk Assessment Algorithms at Sentencing.” I especially like this one for your prediction paradox. Her argument includes the idea that predictions can become recursive: society acts upon the prediction, which changes people's circumstances and behavior, which can cause reality increasingly to resemble what the algorithm predicted. In other words, the prediction isn't simply observing the future—it may participate in creating the future it claims to predict. That's extremely close to the philosophical problem you've identified.Read the open-access paper
5. Stanford Encyclopedia of Philosophy — “Causal Determinism.” This is my wildcard, and I'd definitely include it because your group is Politics and Philosophy, not the Sacramento AI Policy Committee. The wonderful thought experiment here is Laplace's Demon. Imagine an intelligence that knew the position and forces of everything in the universe with sufficient precision. Could it calculate the entire future—including every decision you will ever make? AI gives that centuries-old philosophical thought experiment an eerie new relevance. And there is an important distinction for your group to wrestle with: predictability doesn't necessarily prove determinism. An AI becoming extremely good at predicting Susan doesn't automatically establish that Susan has no free will.Stanford Encyclopedia of Philosophy — Causal Determinism
Sources are optional. Pick the perspective that interests you most and bring one idea from it to the discussion.
Intellectual progression among these readings:
Laplace: Can the future theoretically be known?
↓
Harcourt: If we can predict people, should government act on those predictions?
↓
Schwerzmann: What if acting on predictions changes the people being predicted?
↓
Zuboff: What if corporations discover that predicting—and influencing—people is enormously profitable?
↓
Acemoglu & Johnson: Then the decisive issue becomes who possesses the technology and who has the political power to determine how it is used.2 attendees
Past events
1001
