Who Owns Your Information in the Age of AI?
By John R. Kennedy ·
The difference between having access to your information and having control over it
Somewhere in the world, there's a computer system containing information about you that you've probably never seen. It might know where you've lived, what you've purchased, how much money you earn, or which websites you've visited. Other systems contain information about your health, employment, family, and financial circumstances. Some of that information was provided willingly, some was collected as you went about your daily life, and some may have been inferred from your behavior without you ever being asked.
We've become so accustomed to this arrangement that we rarely give it much thought. We open accounts, install applications, sign agreements, and share information because doing so allows us to accomplish something we want. We need to see a doctor, purchase a product, communicate with friends, or access a service. The information we provide becomes part of the transaction, and we generally move on without considering what happens to it afterward.
For most of the digital age, this arrangement has worked reasonably well from the perspective of the organizations providing those services. Healthcare providers maintain medical records, financial institutions track transactions, employers document our professional activities, and online services collect information necessary to operate their platforms. Each organization has its own reasons for maintaining these records, and many of those reasons are entirely legitimate.
But I've always found something peculiar about the relationship we've developed with our own information. We are the common element connecting all these systems, yet we're often the least equipped to assemble a complete picture of ourselves.
Imagine deciding tomorrow that you wanted to collect every meaningful piece of information about your life. You might begin with your medical history, contacting physicians, hospitals, pharmacies, and insurance companies. From there, you could move to financial institutions, former employers, schools, and government agencies. Some organizations would provide information electronically, while others might require forms, identity verification, or considerable patience.
Eventually, you'd accumulate a collection of documents and records that describe different parts of your life. But even then, you wouldn't necessarily have a complete picture. There might be information you didn't know existed, records containing errors, or conclusions that organizations had drawn about you without your knowledge.
And assembling the information wouldn't necessarily give you control over how the original records were being used.
I think this distinction between access and control will become increasingly important as artificial intelligence changes what we can accomplish with information.
Consider your medical history. In many circumstances, you have legal rights to obtain copies of your records, but that doesn't mean you can determine every way the original information is maintained or used. Healthcare providers have responsibilities to retain certain records, insurers need information to process claims, and researchers may be permitted to use appropriately protected information under particular conditions.
These arrangements reflect the fact that information doesn't fit neatly into our traditional understanding of property.
When we purchase a house, the concept of ownership is relatively straightforward. We understand that the property belongs to us, subject to laws and other obligations. Information is different because multiple people and organizations can have legitimate interests in the same record.
A physician creates a diagnosis, but the diagnosis describes the patient's health. The healthcare organization maintains the record, while an insurance company may need some of its contents to process a claim. Each party has different rights and responsibilities, and those rights don't necessarily amount to exclusive ownership.
This complexity has existed for decades, but AI introduces an entirely new dimension.
Until recently, much of the information collected about us remained relatively isolated. A retailer understood our purchasing habits, while a healthcare provider knew our medical history. A financial institution had insight into our spending, while an employer maintained records of our professional activities. These collections were valuable, but connecting them often required considerable effort.
AI is changing that.
Information that once required hours of human analysis can now be examined much more efficiently. Patterns that were difficult to recognize may become apparent when information from different sources is considered together. The ability to connect these fragments could fundamentally change how we manage our lives.
Imagine having an AI assistant that understands your medical history, medications, insurance coverage, and personal schedule. Instead of repeatedly explaining your circumstances to different healthcare providers, you could authorize the assistant to assemble relevant information and help coordinate your care.
Perhaps it notices that a medication prescribed by one physician could interact with something recommended by another. It might identify a preventive screening you've overlooked or help you understand why an insurance claim was denied. The assistant wouldn't need to replace your physician to provide meaningful value. Simply helping you organize and understand the information could make navigating healthcare considerably easier.
I find that possibility exciting because it illustrates how artificial intelligence could return some of the value of our information directly to us.
But it also reveals a complication.
For the assistant to provide those benefits, it would need access to some of your most personal information. It might also need to communicate with healthcare providers, pharmacies, insurance companies, and other services. Each interaction would involve decisions about what information was necessary and how it could be used.
Today, many of those decisions are governed by agreements and technical arrangements that ordinary people have little opportunity to understand. We frequently authorize broad access because it's the easiest way to obtain a service, even when the service requires only a small portion of the information we're making available.
I don't think most people consciously choose to surrender control over their information. More often, they accept the available arrangement because there isn't a practical alternative.
And that brings me to something I've been thinking about for a long time.
We tend to associate assets with things that have an obvious financial value. A house, an investment account, a vehicle, or a business. These are things we can identify, protect, transfer, and sometimes sell. We understand their importance because their value is relatively easy to recognize.
Information is different.
Your medical history has value because it can help you make better healthcare decisions. Your employment history reflects experience and qualifications accumulated over a lifetime. Family photographs preserve moments that may be impossible to recreate. Even your personal preferences, habits, and the knowledge you've acquired through experience can become valuable when they help you accomplish something meaningful.
None of these assets necessarily needs a market price to have value.
In fact, I suspect some of our most valuable information is information we'd never consider selling.
A photograph of a deceased family member may have little commercial value, but to the people who loved that person, it could be irreplaceable. A detailed medical history might be of limited interest to most organizations, yet become extraordinarily important when a physician is trying to understand an unfamiliar condition.
The value of information often depends on the circumstances in which it becomes useful.
Artificial intelligence has the potential to make that usefulness much more apparent.
Consider how much knowledge we accumulate throughout our lives. We learn from mistakes, develop professional expertise, establish relationships, and gradually build an understanding of the world around us. Some of that knowledge is documented, but much of it remains scattered among emails, photographs, conversations, personal notes, and our own memories.
For most of history, preserving and organizing that information required considerable effort. Even when we successfully recorded it, finding the right piece of information at the right moment could be difficult.
AI could change that relationship by helping us connect what we already know with what we're trying to accomplish.
A business owner might use years of correspondence and operational records to understand how a particular decision was made. A parent could preserve family stories in a form that future generations could explore. Someone managing a complicated medical condition might develop a more complete understanding of their health by connecting records that were previously scattered across different providers.
These possibilities suggest that the value of our information may increase as our ability to use it improves.
Yet much of the information necessary to realize that value remains distributed among organizations whose interests don't always align with our own.
I don't believe the answer is to prevent information from being shared. Modern society depends on the responsible exchange of information, and artificial intelligence could make those exchanges considerably more beneficial.
Instead, I think we should reconsider the relationship between the people information describes and the organizations that collect and maintain it.
Imagine a future in which you maintain a trusted collection of information about yourself. Your healthcare providers would continue keeping the records they're required to maintain, financial institutions would operate their systems, and employers would retain information necessary to run their organizations. But you would also have a practical way to assemble information relevant to your own life, understand its origins, correct inaccuracies where possible, and authorize specific uses.
An AI assistant helping you prepare for a medical appointment could work with your relevant health history without needing access to your financial accounts. Another assistant helping you apply for a mortgage could use verified financial information without requiring access to your personal photographs or private conversations.
The objective wouldn't be to isolate information from the world. It would be to make sharing more intentional, allowing people to benefit from their information without unnecessarily surrendering control over it.
Of course, making this possible would require more than developing sophisticated AI. Information would need to be trustworthy, permissions would need to be understandable, and organizations would need reliable ways to exchange information while respecting the rights of everyone involved.
There would also be situations in which one person's information overlaps with another's. A family photograph, medical history, or business correspondence may involve several people with legitimate interests in how it is used. Greater individual control cannot mean ignoring those shared responsibilities.
These challenges are significant, but I don't believe they're reasons to abandon the idea. They're reasons to begin thinking more carefully about how we design the systems that will increasingly manage our information.
There's another development that makes this conversation particularly timely.
As artificial intelligence becomes more personalized, it may begin accumulating something that doesn't exist in quite the same form anywhere else: an understanding of you.
Think about the conversations you might have with an AI assistant over several years. It could learn about your projects, interests, preferences, and the way you approach problems. It might understand the history behind decisions you've made or recognize connections between different parts of your life.
Some of that understanding would come from information you deliberately provided. Other aspects might emerge from patterns in your interactions.
Over time, this accumulated context could become extraordinarily useful.
An assistant that understands your circumstances wouldn't need you to explain everything from the beginning each time you asked for help. It could recognize the significance of information that might seem unimportant to someone encountering it for the first time.
But imagine spending five years developing that relationship and then deciding to use a different AI service.
You might be able to export your documents and conversation history, but would that preserve the understanding accumulated through those interactions? Could another system reconstruct the context that made the original assistant useful? Would you even know what information had been retained or inferred about you?
We've encountered versions of this problem when moving between software platforms, email providers, and social networks. But the stakes may become considerably greater when we're no longer transferring only files and account records.
We may also be trying to preserve accumulated understanding.
I find that possibility especially interesting because it suggests that our digital assets are becoming more than collections of information. They may increasingly include the relationships between pieces of information and the context that gives them meaning.
And if that context becomes essential to how we work, make decisions, and manage our lives, losing access to it could be far more consequential than losing access to an individual application.
I sometimes wonder whether future generations will look back at our current approach to personal information and find it difficult to understand. Not because every piece of information should have a price or because people should be able to exercise absolute ownership over every record that mentions them.
Rather, because we may be living through a period in which the value of personal information is increasing dramatically while the people it describes have relatively little practical ability to manage that value for themselves.
Artificial intelligence gives us an opportunity to reconsider that arrangement.
We can continue building systems in which information is collected, analyzed, and controlled primarily by the organizations providing services. Or we can begin exploring ways to give individuals a more meaningful role in determining how their information is organized, protected, and used.
I suspect the future will involve elements of both approaches. Organizations will continue to have legitimate responsibilities, and people will continue to benefit from services that require sharing information.
But I hope we begin to recognize that the person at the center of all those records deserves more than occasional access to fragments of their own life.
In a world where artificial intelligence can make our information more valuable than ever, perhaps it's time we become more than just the subjects of that information.
Continue reading: The 10 Levels of AI Thinking: Where Are You?