AI and the Questions We're Afraid to Ask: Why Autonomy, Privacy, and Freedom Matter

The best thinkers I know aren't afraid to ask hard questions. But bold ideas don't show up fully formed. They begin as fragile drafts: uncertain, exploratory, not yet ready to be seen.
TL;DR
- When every question becomes a record, people stop asking the ones that matter most. That is the chilling effect.
- The court order in The New York Times v. OpenAI, requiring the preservation of all output data, including deleted conversations, sets a precedent with implications far beyond a single copyright case.
- Private, exploratory, rough-draft thinking is the necessary condition for creative and intellectual growth. Surveillance eliminates those conditions.
- Autonomy is the root of freedom. Without autonomy, no privacy. Without privacy, no freedom. Without freedom, no formation.
- This is not a legal problem. It is a formation problem, and higher education, of all institutions, should be willing to say so out loud.
My dad is a lawyer. Growing up, he used to say something I didn't fully understand until I stepped into my own classroom: "Autonomy is the root of freedom. If you lose autonomy, you lose freedom." He argued that without autonomy, one loses privacy, and without privacy, one loses freedom.
I've been teaching communication law for years. I know how to read a legal argument. But when I read about the court order in The New York Times v. OpenAI, directing OpenAI to preserve all output data, including conversations users had deleted or believed were private, what unsettled me was not the legal mechanics. It was the precedent.
AI privacy and intellectual freedom are not abstract policy concerns. They are the conditions that make genuine thinking possible. And a world where every question becomes a potential exhibit is a world that makes thinking harder in exactly the ways it most needs to be free.
When every question becomes a record, we stop asking the questions that matter most.
PRIVACY WILL BE A CHALLENGE IN AN AI WORLD
AI Privacy and Intellectual Freedom: The Court Order That Unsettles Me
[]()
The case itself begins with a straightforward legal claim. The New York Times is suing OpenAI for copyright infringement. As part of that litigation, a federal judge directed OpenAI to preserve all output data, including conversations users had deleted or considered private. The legal rationale is understandable: courts need evidence, and evidence must be preserved.
But step back from the legal rationale and ask the broader question: what happens when the entire corpus of private, exploratory, often-deleted thinking becomes a preserved record?
AI tools like ChatGPT are not just calculators or search engines. They have become something more, sounding boards, scratchpads for imagination, the place where half-formed ideas go before they are ready to be shared. Increasingly, they are part of how we think.
Asking a question in a private AI conversation is not the same as making a public statement. In the same way, a late-night search for medical symptoms is not a medical opinion. The same way a journal entry is not a confession. The cognitive work of private exploration is categorically different from the outputs of finished thought, and treating the two as equivalent has consequences that extend far beyond any single lawsuit.
WHAT THIS TELLS LEADERS
This is not a controversy limited to tech policy or intellectual property law. It is a question about the cognitive conditions we are building for an entire generation and whether the institutions most responsible for forming thinkers are paying attention to what is being quietly eroded around them.
SAFETY IS ESSENTIAL TO LEARNING
We Don't Think When We're Being Watched
There is a well-documented psychological effect: we behave differently when we believe we are being observed.
In classrooms, I have watched students grow bolder, more willing to ask the messy, uncertain, embarrassing questions when they feel genuinely safe. The learning that occurs in those moments is qualitatively different from that which occurs when students perform for an audience. Private exploration enables the kind of thinking that public performance forecloses.
This is not an abstract principle. It is what every good teacher knows intuitively.
When we begin to treat AI conversations like legal documents, as permanent records that could be subpoenaed, disclosed, or misinterpreted, we introduce the cognitive cost of observation into every interaction. People will self-censor. They will avoid questions that feel risky. They will stop using AI as a scratchpad and start treating it as a presentation surface.
- What will not get asked when people worry someone is watching?
- What will not get explored when curiosity might be misinterpreted?
These are not hypothetical costs. They are the practical toll of a legal framework that treats thought-in-progress as evidence.
[]()
WHAT THIS TELLS LEADERS
The chilling effect is not a theory. It is already happening every time a student hesitates before asking a question that might look incriminating, or every time a faculty member thinks twice before exploring an idea that hasn't been validated yet. The question for educators is not whether this affects learning — it does. The question is what we are willing to do about it.
When the scratchpad becomes a surveillance record, thinking changes.
Most leaders we talk to say the same thing: "I hadn't thought about privacy as a formation issue — I thought it was a legal and IT issue."
You do not need to solve this alone. You need someone who has helped institutions like yours protect what inquiry requires.
[Let's talk about where your institution stands →](/contact)
Here is what protecting intellectual freedom in the age of AI actually requires:
- Explicit privacy language in your institutional AI policy
- Faculty development that names the chilling effect before it shapes behavior
- A public institutional voice in AI data governance conversations
- Curricular spaces designed for genuinely private exploratory thinking
- A partner who has helped institutions make this argument with clarity and conviction
YOU NEED A SAFE SPACE TO CREATE AND INNOVATE
Creativity Requires Privacy
[]()
As a filmmaker, I know this with certainty: great storytelling starts with bad ideas. You have to sift through the wrong turns, the half-formed thoughts, the uncomfortable questions you are embarrassed to have asked. That is the only way to find anything worth keeping. The rough draft is not a lesser version of the work. It is the necessary condition for the work to exist at all.
When you introduce the awareness that your bad ideas might someday be pulled into the spotlight, that your exploratory, deleted, half-formed thoughts are being preserved in a database, you stop being brave enough to have them.
And when you stop being brave enough to have bad ideas, you lose the path to the good ones.
This is not unique to filmmakers or storytellers. It is the condition for any genuine thinking. The scientist who is afraid to follow a counterintuitive hypothesis. The student who will not ask the question that reveals a gap in understanding. The leader who will not admit uncertainty in private before offering clarity in public.
The exploration that knows it is being watched is not fully free. And the imagination that cannot afford to be wrong does not take the risks that matter.
WHAT THIS TELLS LEADERS
The creative and cognitive conditions we protect in our institutions are not secondary concerns. They are the conditions that make education possible. The moment we treat private inquiry as potential evidence, we have already conceded something we cannot easily recover.
THIS IS A HANDS-ON ROLE
AI Privacy and Intellectual Freedom Is a Formation Problem, Not Just a Legal One
Here is the reframe. I am not here to argue the merits of The New York Times v. OpenAI. Copyright protection matters. Courts need evidence. I understand why the order was issued.
But fair does not always mean just. And protecting evidence should not come at the cost of eroding the cognitive rights of millions of people who were using a tool they understood to be private. There must be a better way to balance intellectual property protection and user privacy because that tension is exactly what courts are meant to navigate, between protection and overreach, between law and liberty.
This is not a legal problem. It is a formation problem.
Formation, not information.
My father's maxim was never about data. It was about people, about what we become when we are free to think, to explore, to ask the questions that don't yet have respectable answers.
Autonomy is the root of freedom.
Without autonomy, we lose privacy. Without privacy, we lose freedom. Without freedom, we lose the capacity to form ourselves to become, through inquiry and doubt and the slow work of thought, the people we are trying to be.
The urgency is new. The work is not.
The fight for intellectual freedom, the right to think privately, messily, without surveillance is ancient. AI has made it newly urgent in ways that should concern every institution whose mission is to form thinkers rather than produce credentials.
WHAT THIS TELLS LEADERS
This is a freedom worth protecting. The question is whether higher education, the institution most responsible for the conditions of free inquiry, is willing to say so out loud, in its policies, in its public voice, and in the moments when it has the chance to shape what comes next.
Your institution cannot protect intellectual freedom alone. AI First Responders can.
A single leader, no matter how skilled, cannot be present in every department meeting, every classroom, every moment when a faculty member or staff member needs support. That gap is not a failure of leadership. It is a structural reality.
You do not have to choose between central leadership and distributed capacity. The institutions building AI cultures that last are investing in both.
[Learn about the AI First Responder program →](/ai-first-responder)
The AI First Responder credential trains peer leaders across your institution to fill it:
- Faculty and staff certified in a practical diagnostic framework
- Embedded in departments, not housed in a central office
- Present at the point of need — when the question is real and the stakes are immediate
- A distributed network that extends your AI leadership into the daily fabric of the institution
What Higher Ed Leaders Can Do Right Now
Higher education is uniquely positioned to protect the conditions of free inquiry in the age of AI. That protection requires more than policy language. It requires a stance.
Name the cognitive stakes explicitly in your AI policy. Most institutional AI guidelines address academic integrity and appropriate use. Few address privacy and intellectual freedom directly. Add language that explicitly acknowledges students' and faculty members' right to exploratory, private inquiry and that distinguishes AI tools used for learning from surveillance records.
Have the privacy conversation with students before they self-censor. Do not wait for students to encounter the chilling effect before you name it. Talk openly about what it means to think with AI, what privacy protections currently exist, where the limits are, and why protecting private inquiry matters. Equipping students with this framework is part of preparing them for the world they are actually entering.
Advocate publicly for user privacy in AI governance. Institutional voice matters in policy conversations. Submit comments to regulatory bodies. Engage accreditation discussions. Participate in the emerging debates about AI data retention, user consent, and intellectual property. The institutions that are silent in these conversations are the ones that will live under policies written by others.
Identify AI tools and environments that offer meaningful privacy protections. Work with IT and legal teams to understand the protections available in the tools you are recommending to students and faculty, and design curricular spaces where exploratory, low-stakes thinking can occur without unnecessary surveillance risk.
Move from AI anxiety to AI strategy. The privacy concern is real. It should produce a plan, not paralysis. Institutions that engage proactively with AI governance, data rights, and intellectual freedom will be positioned to protect their communities. Institutions that avoid the conversation will be shaped by decisions made elsewhere
The argument this post makes that AI privacy and intellectual freedom are foundational concerns, not just legal ones, needs to be heard in every institution that believes free inquiry is central to what it does. The AI First Responder credential trains peer leaders to carry this argument into departments, policy conversations, and faculty development sessions: that the conditions AI requires to be used well are the same conditions that great education has always required. If your institution is navigating AI governance and you want to protect what thinking actually needs, this is the work
[Let's chat](/contact)
COMMON QUESTIONS
Frequently Asked Questions
What is the chilling effect in the context of AI?
The chilling effect refers to the tendency to self-censor, restricting thinking, questioning, or exploration when people believe their actions are being observed or could have negative consequences. In the context of AI, the chilling effect occurs when users become reluctant to ask exploratory questions or use AI as a private scratchpad because they fear those interactions could be preserved, disclosed, or misinterpreted. The cognitive cost is real: people stop having the uncertain, half-formed, "bad idea" conversations that are often prerequisites for genuine learning and creativity.
What should higher education institutions do about AI privacy?
Start by naming the cognitive stakes explicitly in your institutional AI policy, not just the academic integrity concerns, but the intellectual freedom concerns. Have the privacy conversation openly with students and faculty before self-censorship becomes the default. Advocate in the policy conversations where AI data governance is being written, and identify tools that offer meaningful privacy protections for the exploratory thinking that education depends on. The institutions that protect intellectual freedom in the age of AI will be the ones that make it a deliberate priority, not an afterthought. If you want help building that argument and that framework at your institution, this is the work.
How does AI privacy affect education and learning?
Learning requires the freedom to ask questions you are uncertain about, to explore ideas that might be wrong, and to think through uncomfortable possibilities in a genuinely low-stakes environment. When AI tools become part of how students think as scratchpads for inquiry before ideas are ready to share, the privacy of those interactions directly affects the quality of learning they enable. A student who worries that an AI conversation could be misinterpreted or disclosed will think more carefully before asking and less freely while exploring. That calculus is the chilling effect applied to the act of learning itself.
What is the relationship between autonomy, privacy, and intellectual freedom?
The most durable model pairs a Chief AI Officer with a distributed peer network — trained colleagues embedded across departments who can provide AI support at the point of need. A single central leader, however capable, cannot be present in every classroom, every meeting, and every moment of uncertainty. A peer network extends that leadership into the daily fabric of the institution. The AI First Responder credential is designed to build exactly that kind of distributed capacity — and the institutions that invest in both central leadership and peer networks are the ones that see lasting change.

Let's talk.
If you are leading an institution through this moment and you are not sure where to start, that is exactly what I do.
[Request a consultation](/book-a-call)
Sarah Gibson is Director of the School of Communication and Faculty Fellow for AI at Lipscomb University and founder of PivotPoint EDU. She helps higher education institutions and nonprofits move from AI anxiety to AI strategy with clarity and confidence. Learn more.
Comment 0 Comments JD