Michelle Finneran Dennedy: Privacy Is the Heartbeat of Humanity | Intelligence at the Frontier 2026

It's Saturday night. Welcome to privacy. Um Thank you so much, makers. Woo! We could cheer for that.

It's Saturday. This is how we are choosing to rock our social lives, guys. This is where we're at in our lives right now. Um thank you guys so much for coming, for staying, for leaving the robotics lab and the taco floor to come and talk about these important issues. I'm so proud to come after tour.

We have saved lives. Privacy has saved lives. Privacy has saved companies. Privacy has bought my house. When we think about privacy when I think about privacy and hopefully from evermore you will think about privacy some people first go to my old boss, Scott McNealy.

Remember Scott McNealy? Remember Sun Microsystems used to be a pretty big deal? Well, about 1999 Scott McNealy said you have zero privacy followed it up with a global [ __ ] you of get over it. He did this at a press conference and it actually was before there was Atlassian Jira. There was a product based on Java called Jira.

And this little product allowed you to take your laptop and send a print code out into the world and remotely print a document. And a reporter came and said, "Isn't this the end of privacy?" And he said, "Ah, privacy's over. Get over it." And he did need it.

Poor little lamb. He was wrong. Now back then I was a patent litigator and when I came to Sun and our CEO had said there's zero privacy, get over it. You can imagine how many people wanting promotions and budget ran towards the privacy issue. Answer?

No. No one did but me. I came to line and said, "Listen, I used to be a patent litigator. The way we roll cuz we're even more exciting on a Saturday night we break things down into claims. And inherently if you patent your your idea, or you get a copyright, or you get a trademark, what you're doing is you're doing a digital hug of ownership around a new and non-obvious idea that you hope will somehow propel humanity forward.

Well, I look at privacy very much that same way. And when I think about privacy, it really is the heartbeat of humanity. It's an This is This is why I said this is a a manifesto for the agentic age. Because now we're no longer just talking about what can human do to human? How can I violate your privacy?

How can I sell you things you don't want to buy? How can I tell a lie about your mortgage or your FICO score so I can steal from you or compromise you or or mistrust you or just sell your stuff to someone else? Now we're talking about machines. Machines acting on machines. So, when we think about privacy and we think about particularly the advanced AI society bringing together a multitude of different types of makers and dreamers and thinkers and policy makers and financiers, getting them together in an advanced AI society, we're getting out in advance of this agentic world where AI really can be used to promote and defend privacy.

And so I ask you, a lot of people are back in the Scott McNealy camp of what is privacy. It's really important before you solve a problem to define the problem. And so when I break down into the claims if as it were of privacy, what is privacy in the first place? Privacy is the authorized processing personal data according to ethics, law, and sustainable commerce. Simple, right?

What happens with authorization when I sit down at a restaurant and I order a coffee? It's kind of my table. I can whistle up and say, "Bare, please." There are some. In the middle of the night if I shower and I sit down at that same table, I'm bringing in energy.

Authorization is contextual. If I just tear up my clothes in the middle of an office, it's typically frowned upon. When you're a year at 60, Really frowned upon. When I'm in my doctor's office, I am paying her for the privilege to see all of this, lucky gal. Context matters, right?

Privacy is sexy, people. Sexy. So, if we you think about privacy as the authorized processing, processing is not a boring, it's not digital waste, it's not digital discard. Processing, if you look at it through the eyes of the objects, which is Russia's Google, it's gone from $2 billion a few years ago to $14 billion in lines in What is that based on? It's based on raw data.

It's based on my data. Do you really think $14 billion the value before running over to duplicate the browser activity Russia? No. They're walking down the same virtual corridor and scraping information about all of our transactions, all of our activities, all of our conversations, and they're making decisions about us. And that's why I keep going back to privacy's connectionality.

Every data set describes human behavior. Even when they anonymize it, all of us are telling a digital story every single day that deserves to be authentic, it deserves to be told in context, and we deserve some modicum of control and agency and proof that we have sustained value. So, what are the potential proofs of control? So, this is one of the tenets of the advanced AI society. What are the proofs of control?

We built a to a couple of years ago called privacy code and we were we were acquired by Abacus Technology where I currently work. But the premise was you could actually take a document, you could take a requirements document, a contract, a policy and feed it into what I'll call an SLM, a small language model of known activities that have to be done and executed. So beyond the checklist of do we have a policy, when we say we're being transparent and we care about privacy, if you look at almost every privacy policy, there's a lot of caring about privacy. It's almost like a Hallmark movie channel. We care about your privacy.

What does that actually break down to mean? Does that mean there are people assigned to it? That's an HR goal. I can measure that as a proof. Does that mean that on every single webpage I have something protecting you from being collected or surveilled outside of your knowledge?

Yes. Does that mean on the back end we spend time and money into deletion and segregation and harvesting actual quality relevant information? All of these are proofs of control that are owned by different types of specialties. So now in an agentic world we have both the compute power and the ability to break down a proof of control and actually have what we called the privacy engineers manifesto, a privacy component. What are those cellular level, those nano level rules that we can now orchestrate and score?

So now we're starting to measure and now we have a proof of control and if it doesn't exist, we don't do business. If it doesn't exist, we don't extend that funding. If it doesn't exist, we make it exist and that's where a maker space comes into play. We are building the signals that will help future makers and future users of this new agentic world know when they're drifting off course. It's not going to happen sua sponte.

You're not going to hire all the lawyers to write policies around it after the fact. We're building in signal today. So, understanding what those proofs of control matters in planning, in governing, in holding accountable, and shining that signal for future navigators. So, we come to what I mostly and lastly call fair and smart fair and mighty name. Success is a very simple formula.

Success equals data value greater than data risk. Simple, right? Column let me have this what is value and to whom? And how do we keep sustainably looking to proving that value? And when we're talking about data value, of course it changes over time.

Of course it changes from perspective. Of course it changes in the collective versus the singular. So, not everyone's ethical are is necessarily the same, but you have the right to at least understand what it is. So, understanding what is data value also depends on what we value. As Peter Drucker, the great management guru, said, we can we manage what we measure.

What fairness ought implies that we also protect that which we treasure. So, you can know what value is by needing into the metrics that create purpose, value, integrity, quality, safety. And those things are balanced into a collective algorithm, and that can be turned into remediation. And that's how you balance against risks instead of the over rotation where we are today of will I get me sued? Will I get caught?

Will I get embarrassed? All of those things are critical, but if you're only measuring data value, how can you know whether risk is acceptable? So, finally the mission. I actually wrote this down. The future of AI will not be decided what machines you can do.

Machines can do a lot. The future will be what we builders choose to protect. So as we start to build in now, assume that we will be wrong. Assume that that thing that you thought was going to protect all of those people will find someone will find a tunnel in. How will you know when that tunnel in has happened?

Assume that this collection of information that know is valuable will start to decay. We find this particularly in the mature years of your career, may say. Or sometimes I like, "Ah, I got the perfect guy. He's going to help fund my next company." That guy's retired.

It's happening more and more. I mean, that's a That's the best case. He's retired. He might go somewhere else. Understanding the data about who the means changes its complexity as quickly as humans change their complexity.

So understanding and applying it sociologically, my friend Trisha, making sure that we're coding in the right things, but also making sure we have the capacity for humility. Proofs of control and navigational signals, they're not religion. They're not special. We're not that special. Sorry, that's a bummer for a Saturday night.

What is special is our collective ability to decide what is special and valuable and important and then to act. Act on that goal. Build today. Keep communicating about what you think the answer might be. And then we end up at a really beautiful place.

And it actually turns privacy into a pretty damn sexy ride. I've I've doing it for almost 30 years now. I'm having a ball. It's the heart of where we're going next with a gentle AI and it's the heart of what we're going to do with cubits next year when we have the same conversation in a quantum environment. And I invite you all into this beautiful, messy, massively human parade.

That's what I got.