The AI Future Nobody's Talking About: Trust as Infrastructure

Full Episode Transcript

This is the full recorded conversation from The Human Protocol Podcast, published by The Human Protocol Group. © 2026 The Human Protocol Group. Please link to and attribute the original episode when referencing this conversation.

THE HUMAN PROTOCOL (00:02.039)
What if data protection isn't a barrier, but the blueprint for AI? Compliance shouldn't be the enemy of innovation, it should be the condition for it. Welcome to The Human Protocol, the podcast about staying human while building the future. I am your host, Michael Salomon. Today's conversation, it is about something most companies and governments and organizations get completely wrong.

They see compliance as a friction. They see data protection as blocker. They see governance as something to deal with later. But what if that's exactly backwards? My guest today is James Robson, former data protection officer for the UK Labour Party and someone who had spent the last decade building real-world applications of GDPR and ethical data sharing.

James, welcome to the Human Protocol. It's a pleasure to have you with us today.

James Robson (01:05.673)
Michael, thank you so much for having me on. It's a complete pleasure.

THE HUMAN PROTOCOL (01:09.032)
Absolutely. I appreciate you taking the time to be with us today and to talk about this important topic that we're going to be covering today, right, which is related to data protection, especially in times of AI. And to kick things off, I would like to ask you, James, to give us kind of an introduction of who James is, where have you been, and where are you heading next?

James Robson (01:38.923)
Thank you very much for the opportunity to tell people about my good self. I've been in data protection for the last decade, like you said, and I started out as one of the early GDPR, general data protection regulation consultants in 2016, early 2017, and over a few years there did these sort of gap analysis. Where is an organization to where it needs to get to become

client to that regulation and picked up lot of skills, a of awareness, a lot of knowledge, jumped into a large psychometric testing firm that was in 120 countries and they were doing 50 billion data point AWS migration from in-house servers to AWS. And then I jumped into research, but not me doing the research, it was...

me doing the data protection for a set of research organisations concurrently, all funded by UK government and that led really nicely into me

being able to become the data protection officer for the UK Labour Party in the middle of 2023, so a year before the UK general election and also the year after. And it was kind of towards the end of last year where I wanted to follow my passion for data sharing ecosystems for societal benefit specifically, where as you said in your opening,

Data governance and data protection can be the enabler for that societal benefit and use of data to help people for good, data protection for good, data being able to save lives. Where people think it would be unethical or unreasonable, even illegal to share data for those things, I see a way through that because of my depth of knowledge in that field to go, well, it's not a don't do, it's a how to.

James Robson (03:45.291)
How would you do that in the right way to then get to where you need to be? Because, you know, I really like the analogy of you don't have brakes on a car to go slower. You have brakes on a car to go faster. Without brakes, and you go as fast as you can, you crash and you die. If you don't have brakes on how you use your data, it gets everywhere. People get hurt. You get sued.

and your company tanks or your organization folds because trust has to be built into that. And that's why governance is important. That's my passion. That's why I love it. I've trained as a yoga teacher in the past. I trained as a filmmaker in the past as well. So I've done a bunch of things, but this one I've really got my teeth stuck into and I love this one.

THE HUMAN PROTOCOL (04:35.726)
Awesome, thanks for that. I know you and I, both don't see data protection as really a limitation. We see it as a manifesto for really innovation, right? And that's what we are gonna try to explore today and what that actually means for AI, for society, and really for trust. And I want to get back a little bit on what you just said at the beginning, because that was one of the things that I wanted to ask you first. From what I know about you, you started in film,

James Robson (04:45.237)
Yeah.

THE HUMAN PROTOCOL (05:05.821)
storytelling, right? And I'm curious if you can share with us how did that shape how you think about systems and data today.

James Robson (05:16.073)
Yeah, it's knowing the outcome and what you're aiming towards and then how to tell people about that because, know, it all gets simplified for those that haven't read the legislation into black and white, yes or no, we don't think we can. And then what I'm able to do is use that storytelling techniques to say, well,

Think of this scenario where you could have this outcome. And I'm not talking about data anymore. I'm saying that maybe somebody who is homeless, you may be able to find out where they are, at what point they become homeless to reduce the...

reduce the possibility of them becoming homeless in the first place because you are hitting those data points in the right way and you're curating that data in the right way to go, well, you can look at it if it's for the right benefit and outcome and would be acceptable to the person whose data is being used because, you know, I'm going to talk to you about one of the AI projects I've been doing the governance for, which is trying to do exactly that thing around homelessness. you know, using the storytelling has been

absolutely instrumental because not only is it kind of how do you have the know the story arc to get people interested but then you also need to tell it in a way that is super interesting and you can get other people fired up about it so you know all those all that skill set has to be in there and I've used that well not too badly in my career so far yeah

THE HUMAN PROTOCOL (06:55.854)
I like that analogy and how the storytelling can help you connect more on this topic with people, right, by making sure that they understand where they are in that journey and how we can connect the dots and getting from point A to point B based on the outcomes that you want to achieve. I really see how that connection is done. So I'm sure it's storytelling and that type of skill.

it's really always gonna be a plus, right? Especially in the profession that you are right now. And what really pulls you into GDPR and data protection, James?

James Robson (07:38.923)
So I think the story comes from after I'd been to film school. I made a 10, pardon me, a 15 part internet drama series back in...

2000 and what were we five no pardon me 2009 it was released in 2010 and I got to the end of that and realized you know what I've done that bucket list tick I don't think I want to do that weirdly it's like it seems like every every you know young lad or ladies dream you know become a filmmaker you know build a portfolio and I got to the end of it and I was burnt out I was broke

no money at all, living back home with my parents, almost 30, and I'm like, this is not for me. And I just didn't enjoy it as much as I thought I would when I put a big project together. So I changed direction. And in fact, I got into IT security sales as a telemarketer, a very old telemarketer, but because of my experience in storytelling and, you know,

There are sales skills if you like, being able to speak well and have the interpersonal skills. I did quite well there and realised that it's quite a skill to be able to pass on complex information to people in a concise, precise way and then train them to be able to perform it themselves in an interview, in a meeting, in a briefing, in a proposal and then be able to sell something.

THE HUMAN PROTOCOL (09:15.871)
Right.

James Robson (09:16.111)
And it was during kind of a few years of that, boss that I had in one of the jobs during that little period said to me, there's this thing called the EU GDPR coming. I think your skillset might work for that. Do you want to have a look at it and tell me what you think? And interestingly, I was coming to the end of my sales career. I wasn't enjoying it. And I looked at that and thought, you know what? I love to learn. And that allowed me to get inside the mind and the business.

of these organizations I'm put in front of to go, okay, where did your data come in? Where does it go? And when do you delete it and why? And you have the framework.

of every organization basically being the same and then the nuance of what they do, which is innovative and above and different and like, this is amazing. And I learned everything from e-commerce to healthcare to dating agencies to, to nightclubs, to hotels, to, you know, every single sector. And I had a two year learning curve of doing kind of a gap analysis for those companies, once every three days for a couple of years. And that

really stood me in good stead to almost have a bit of a superpower to walk into an organization and go like, yeah, I think I know how the data flows here and how you work as an organization. And it was amazing. And it's just taken me on the most phenomenal journey. Two and a bit years as a DPO for the UK Labour Party during a general election. I mean, it's second to none experience and has been so fun.

THE HUMAN PROTOCOL (10:51.234)
Right.

You know, and I totally agree with you. I think it's really a superpower because when you really understand the data, you kind of understand the system, right? Data is really the only function that touches every part of any organization. And that is why it's so important. A system without good data, really, it doesn't exist, right? Data is at the core of everything. All our systems, our processes without data, we really don't have anything.

there. And when did you realize that really data was the map of how everything actually works? When that coming to fruition to you?

James Robson (11:37.286)
Very, very quickly. think within the first few weeks of beginning to meet those organizations, I mean, I was literally thrown in the deep end. The first two weeks of becoming a consultant was my training. The third week I was being flown to Gibraltar to meet one of the gambling firms out there and do a gap analysis. And you very quickly realize that, well, organizations don't survive without personal data, without...

contact with another human being to run themselves, whether it's kind of transacting money from them.

and how do you keep that secure and what are the mechanisms there to actually just having a database to market to them and send a newsletter or keep them updated or service messages. And then you go, well, what is health data? Let's go down that route. And you've got kind of the NHS data framework that you have to align to because that data is quite valuable, could be sold quite easily.

could be marketed to by nefarious threat actors, who are interested in trying to sell the latest pharmaceuticals if they're not allowed to have that data or sell that data for whatever purposes. it's just lifeblood really of organizations and it runs through the skin of the organization. And if you've got a cut and you are leaking that data, that lifeblood, you are basically dying as an organization.

you are losing trust for one, you may be fined by a regulator or be taken to court or sued by a data subject, by an individual because they, yeah, yeah, yeah. Reputationally, it's disastrous. Yeah, completely. All those things. So, you know, very quickly, you get to hear the phrases, data is the new oil.

THE HUMAN PROTOCOL (13:14.882)
Yeah.

on lose trust, right, on the organization. Absolutely.

James Robson (13:36.529)
or data is your money. And I think data probably is your money as well as money and currency now. But what is it? It's lifeblood now. I completely agree with that.

THE HUMAN PROTOCOL (13:48.29)
Yeah. Yeah. Totally, totally. James, let's talk about specifically GDPR, right? There is a narrative, especially in the US in some circles, that GDPR is seen as something that slows everything down, right? That it kills innovation. And I know that you disagree with it, of course.

But why do you think that's the case? Why do you think, especially in the case of US, in some circles, we see GDPR as going in the opposite direction of innovation, for example?

James Robson (14:30.569)
Yeah, so, you know, it's really talked about in the EU and the UK that the difference between data protection regulation between Europe and the US is the US is consumer and customer focused as a legislation, whereas Europe, EU, UK is human rights focused.

And you know, there is the CCPA, the California Consumer Protection Act, which actually, if you compare GDPR and that, there's huge comparisons in there as well. But it's just the open source, open sharing model that feels like without a federated law to protect people and protect their data, it feels like the Wild West out there, you know.

and what the, do disagree with the fact that, you know, GDPR slows things down. think I used the analogy in my opener. mean, what, what if you put data into an LLM that you think is going to make your life more efficient and AI chat bot or something, and you're using other people's personal data in there. And then they're able to find that personal data by,

prompting the same system in another location, maybe in the world, and then finding out that information about you because of that original prompt. I mean, you'd likely be pretty annoyed about that, especially if it's something that you didn't want other people to know about, if you wanted to keep that to yourself. That's pretty bad. Like I said, if you lose it, you lose trust with people.

They don't want to work with you. They will lose customers. You will lose the ability to grow as an organization. People will move away. And one thing I wanted to compare this to is actually, you know, the name of the podcast that we're on, The Human Protocol. I mean, at the moment, there isn't the individual, I guess, self sovereignty ability to go...

James Robson (16:51.035)
I want you to use my data in this way to the standards that I hold for myself. It's lopsided because not just kind of the US EU argument, but in general, kind of across the board.

I mean, everything is weighted in favor of the organization that is using the data and maybe using the data to gain money from you. Let's talk about, you know, MetaPixels and tracking technologies as well that is being sold. You know, Google Analytics used by lots of companies, if not most, that have a website and it is being sold to you.

and it is personal data that is being used and sold so Google can generate revenue. you know, it's the world West generally. think we're, you know, I don't really want to single out the U.S. as being totally bad and totally wrong. It's just scary to us where we see GDPR actually protecting people. And there is the consideration that maybe the U.S. is not protecting people and isn't giving the...

ability for people to make choices about their data because it's just it's taken, it's used, it's potentially abused. They make money out of it. And I don't want that to happen. I'd prefer to have my own human protocol to say, no, I don't want that to happen to my data. And I want it controlled in a certain way. And I want a digital feedback from you to tell me that you are doing it in the right way. And I'm happy about that. And I keep myself sovereignty around the uses of my data.

That doesn't really happen anyway, let alone US versus GDPR.

THE HUMAN PROTOCOL (18:36.718)
Right. Right. that's what you are talking right now. makes me think, what can we do to protect ourselves, right? Especially in times of AI, especially when we don't have any control of what these companies are doing with our data. And especially in the US, regulation is not there to really protect us as it should, right?

So what is it for the humans on this side, right, that are gonna be listening and watching these epics and say, okay, but how can we protect ourselves if we are using child GPT and these LLMs and even social media we have been using forever? Are there any tips or any suggestions on things that we should be looking for on how we can protect ourselves a little bit better?

James Robson (19:25.769)
Yeah, sure. I mean, you know, with having the GDPR in Europe and the UK, it is meant to be built into systems anyway, but talking to maybe an audience out in the US, things like, you know, use an ad blocker, definitely use a password locker as well. So one of those things that stores it outside of the browser specifically, reject all cookies, every

where you can, reject the tracking mechanisms via your browser, do that. Without the GDPR in place, you can't easily request a copy of the data from all the organizations that you have where your data is hold.

THE HUMAN PROTOCOL (20:11.51)
We can or we cannot.

James Robson (20:13.641)
You cannot. You cannot. I'm sure you can actually ask them and there's probably some legislations out there that will allow you to request that. But it's not federated. It's not all states. You wouldn't be able to do that in all states. In California, yes, you can.

THE HUMAN PROTOCOL (20:26.638)
So GDPR forces company that if we ask for all our data, they need to provide it to us.

James Robson (20:35.115)
correct within 30 days as well. There's a couple of stipulations there where it can be extended to 90 days, but you know, it's not a great idea. The regulator doesn't like that so much if you try and extend. Again, the CCPA in California, that's a 45 day window for replying with all the data.

that organization holds on you. And then you get into the realms of what should be redacted, what is their data, what is not your data to be sent back to you and all those questions pop up.

THE HUMAN PROTOCOL (21:10.086)
So do I hear you correctly without regulation in place or any law, it's really up to those companies to determine if they provide our data or not or what they're to even provide if we request our data from them.

James Robson (21:24.777)
Yeah, yeah, absolutely. I'm sure some of those organizations will realize there's no legislation forcing them to do anything based on your request. Of course, you're likely to be slightly annoyed if you don't hear back from them and you've asked them a question. So you may no longer become a customer. there may be, you know, they'd like to keep their customers happy. you know, there are are kind of balances to be had.

by organizations if you want to keep your reputation. I think a lot of it is just based on a little technical handshake, a bit of wording that they are trustworthy on a website, and then it comes down to, how good are their actual processes in the background? And would they take any action if they receive one of those types of emails?

THE HUMAN PROTOCOL (22:15.02)
Yeah, always good stuff and fascinating topic to talk about all these kids. For most people, regular end users that are using all these apps and all these LLMs, people are not thinking about all these, right? They're just using applications. But it's a lot of data that you are inserting in, even about yourself, right? That this becomes data that some of or most of these companies, can sell it. And you cannot do anything about it, right? So.

So I want to talk next about really data for public good, right? What is actually possible when we humans, you know, we work for a greater purpose and trying to achieve better outcome for humanity. And I would like to make this a tangible not only theory, what can data actually do when it's used right? So I want to start by asking you, James, if you can talk.

to us a little bit about the work that you have done with research data and the Office for National Statistics.

James Robson (23:21.322)
Yeah, sure. So I was lucky enough to become the first person to put...

to put non-governmental research data into the UK Office for National Statistics so that that is the most well-known centralised government database for research to be conducted. So second every research on that data. And usually it's big governmental datasets that get in there to be researched. And so I put non-governmental data in there and it was a children's social care data. And the value of that really comes down to

Well, if you can put it in there and maybe do a secondary analysis up against a national pupil database for education, you can work out maybe in the social care sector, does having something like a social worker in a school over a long period of time actually improve the outcomes of children in social care and their educational outcomes? Does it improve their statistics?

their achievements. And then if you do that over the longer term, you could then go, well, let's make this a 10 year or 20 year study and say, well, do they end up claiming benefit? Do they end up homeless? Do they end up what are the markers that you see in these people's lives that actually will, if you step in at the right place, turn them on a different trajectory.

And I've got friends that are working on ideas around proactive contact. So the idea that maybe there's a problem gambler and there's a flag within the gambling shared data ecosystem, which does exist for protecting families from problem parents that are gambling away all their money.

James Robson (25:19.278)
from any gambling institution. Well, what if there was a flag at the right point to then reach back out to that individual and say, we think you're going to be having a problem with meeting rent or causing a problem to your family, we're going to cut you off from all gambling institutions and things.

know, pulling these data sets together in the right way. I'm not saying any gambling data is going to the UNS, it's not, but it's just the idea that you can then go, what are the life markers? And if you can put them into these secure research environments, highly secure, with the right technology in the right way, you can start to work out the trajectory of people's lives.

THE HUMAN PROTOCOL (25:57.166)
Bye.

James Robson (26:04.222)
and where they go off piste, if you like, for one of a better phrase, to then step in and point them in the right direction or put something in front of them.

THE HUMAN PROTOCOL (26:15.637)
I would like to play a little bit of Davies Advocate right now with you. Because, yes, what you're saying makes a lot of sense. We probably have data for almost every human problem that we know. And what happens sometimes is that we're just not connecting the data. And that's where the magic happens. But.

James Robson (26:20.224)
Yeah, yeah.

THE HUMAN PROTOCOL (26:41.516)
The accumulation of data and power in the hands of, let's say, government may scare some people and some people that are fighting for individuals, right? Because if government has all this data and they know so much about us, are you kind of afraid that it becomes out of control as far as the access that government and agency has?

about what we do or what we don't do.

James Robson (27:13.58)
I think there's a technical solution to your answer because it's a fair assessment that if you put all of this data connected in this way into the hands of what could end up becoming fascist or totalitarian or some other kind of extremely detrimental government structure, they could use it against their...

their own people. I agree that there is a possibility for that to happen, but there are technical measures and means that you can do that. Let's look at end-to-end encryption where even the organization that has built the systems aren't able to access that data.

There's the argument around, well, is WhatsApp actually end to end or is it now being shared with Instagram and other meta products probably? And I think they changed the privacy policy around that recently. Whereas then you've got something like a signal where you have more trust because even signal isn't able to look at your messages because of the levels of encryption. And then you have the logging mechanisms around who has looked at what and when and metadata and what they have changed.

and why and you build that in so it's fundamentally embedded into the technical processes for being able to use that data or see that data in the first place. So you know who is doing it, you know why they're doing it, and you know when they're doing it and the length of time they're doing whatever they're meant to do. Stick that onto something like a blockchain immutable ledger technology where you can see every transaction. Yes, I know they're not completely secure for anyone.

insecurity goes like, there's these problems. I get it. There are problems, but generally, you know, they are difficult to hack and crack and break. And if you do that in the right way, you then have these enormous sets of databases with immutable ledges around the uses of them for then use in the right way. And you can map and track and you have these open institutions that are doing this watchful oversight of

James Robson (29:23.137)
the uses of the data in the right way. you know, containerized environments are out there, trusted research environments. There's a system called Five Safes in the UK. So you have, you know, safe data going into it, safe people accessing it, safe locations for access, maybe a pod or a booth, like I'm in a booth here that may actually be air gapped from the internet and you only have access to the

to the secure system as well. And then there's safe outputs. So the safe output is that data is almost held in this airlock sort of container and not allowed to leave that airlock unless it hits certain parameters, has had certain oversight and sign-offs in the right way. So you still have the, call it human in the loop or human on the loop in some way built into those as well.

You know, still mixing the technology with the human and very, very functional design to maintain what is known as functional anonymity around that data, I think is absolutely possible. It's just whether somebody has the courage to set up such systems, run those systems and be scarily transparent about how they do it and not worried about it, you know, all the code being on GitHub.

or Datadog or wherever it is to say, well, I'm completely transparent with how we're running this and there are no holes. Maybe that's a solution.

THE HUMAN PROTOCOL (31:02.594)
Got it, got it. Thanks for sharing that. And just to pick you up on the courage part that you were just mentioning, and especially on this topic of homelessness, right? As you know, we have some cities in the US where homelessness is also a big problem. So I'm curious to see if there's anything that you can share with us that we can learn from what you have seen on how data can help prevent things like, for example,

this issue of homelessness.

James Robson (31:36.162)
Yeah, so that specific issue is really interesting because only a couple of days ago, King's College London in collaboration with an organisation I'm the data protection officer for, called the Centre for Homelessness Impact, they published a study and finding about what they called personal grants. And that was taking a set of homelessness experienced people or people experiencing homelessness and giving them £2,000.

without any prerequisites around how you're to use it, without any judgements, without any rules, just when here it is and we'd like to track you and ask you some questions around how you decide to use that. And the overall outcome was far more positive than you may have thought and people did pull themselves out of homelessness and out of those cycles.

THE HUMAN PROTOCOL (32:31.416)
So basically, the exchange was I give you 2,000 pounds for the government to be able to track your next steps and see what you're heading next, basically, and gather data from it.

James Robson (32:49.132)
Yeah, so it's not necessarily the government tracking a homeless person. was research institutions who were... Yeah, private institutions. So it wasn't around going, we're going to track homeless people and see what happens. It was more around, well...

THE HUMAN PROTOCOL (32:57.952)
Okay. It's a private institution.

THE HUMAN PROTOCOL (33:07.682)
Mm-hmm.

James Robson (33:10.828)
does giving a homeless person £2,000 mean they go off and buy drugs with it and do things that we think they shouldn't do with it? Or do they actually use it to better their lives and then track them over a period of time? Just asking them sets of questions. I think it was just questionnaires they were filling in over a period of time. And then the researchers were able to use other actually government-located data sets around, you know,

this demographic of individuals, what that image of the demographic, not the people themselves, what the image looks like, what are they like, what their markers, their demographic markers, have a look at the government located data sets and go, well, what's the difference between these, because we did this over a period of time, and these over here, when they didn't get that, and what are the markers between them? So.

THE HUMAN PROTOCOL (33:47.264)
Right. Right.

James Robson (34:07.226)
And it came out positive, it came out really, really positive. And that's just, you know, being able to gain the trust of communities, community groups, homeless, experienced individuals, homeless people, to go on that journey with these researchers, accredited institutional researchers, not government again, it's not trying to reach in there. And it had a positive outcome. So that's all data.

THE HUMAN PROTOCOL (34:33.443)
Right.

James Robson (34:36.46)
that is going, let's test this, let's look at it. And now there's a call from that paper to go, well, the government should probably put money into this and it could help a lot of people.

THE HUMAN PROTOCOL (34:38.893)
Right.

THE HUMAN PROTOCOL (34:45.806)
Yeah, again, back to it's all data, right? And it's really fascinating what you can do when you use the data for a better purpose and you start connecting the dots and really diving into what make people take the next steps and how we can prevent it, right? So I think that kind of the magical part of it.

James Robson (34:50.403)
Yeah.

THE HUMAN PROTOCOL (35:10.324)
If we get to that point, I would say, and I would agree, that's a good use of data in a way.

James Robson (35:17.292)
Yeah, I mean, it's almost a little bit of a dream of mine. And I think a lot of people have possibly had this dream where how do you map all the aces in society and then work out

how to address that and preempt any potential outcomes because of that. And ACEs, mean, adverse childhood experiences. And, you know, there's enough literature out there to say if you have this particular type of ACE, then you may have this particular type of trigger mechanism or outcome because of this trauma that you've experienced. And it may be triggered because of X, Y, Z reasons.

And so education is often the way people end up escaping their trauma or being able to integrate with their trauma enough to be able to live their lives without living through their trauma. But wouldn't it be amazing if you could grab an awareness of that experience early enough and then have, you know, the teachers, parents, the guardians, the individuals that help bring up a child.

know what they're tackling with the child and then not pass judgment on the reactions that child has because they're living through that trauma in some way and integrating the trauma incorrectly rather than in a way that's beneficial to them for their entire lives. And that's data as well. That's data on who we are. know, that is our human protocol really. That is our programming. And oftentimes when we able to escape our programming, we can live freer, happier.

THE HUMAN PROTOCOL (36:51.341)
might.

James Robson (36:56.994)
more contented lives, just to be relaxed in our skin, in the moment that we're actually in, and then connect with people a bit better and the world, and you end up building a whole society of people that are just happier, happier to be around others, and more accepting of other viewpoints and peoples and everything really. So that's a dream, yeah.

THE HUMAN PROTOCOL (37:18.956)
Yeah, that's I totally agree. That's a dream. That's my dream. That should be all our dreams, right? And I wish we can we can we can head there or do everything we can to get into heading to that direction. But let's also, you know, be real. There are are bad actors out there. And and I want to to to get and pick your brain on your soul next on on what really what worries you lately. Right.

James Robson (37:24.45)
Hahaha!

THE HUMAN PROTOCOL (37:48.756)
in about all these new words that we're heading into AI, especially when you talk about, I think you have talked about eco chambers becoming full containment systems, right? And I'm curious if you can give us an overview of what you mean by that.

James Robson (38:07.084)
Yeah, I am concerned about echo chambers, you as an individual being locked into

only the specific point in time information that you're displaying to an algorithm that you're enjoying taking a quick look at and the algorithm is always trying to find you what what are you spending more time on what are you clicking more on and if you're in a growth period in your mind an experiential period going i'm gonna i'm gonna look at this i'm gonna look at maybe cartels in columbia i'm gonna look at

heroin trade in Iraq, I'm going to look at how Russia developed through the Tsars and the Gulag. I'm going to look at, you know, I don't know why I'm coming up with horrendous things, but you know, that's what's popping to mind quickly. But then the algorithm goes, you like that? Great. We're going to give you a historical documentation, maybe around the Gulags and Stalin and Russia and how that develops. And we're going to feed that more.

and more and more to you before you know it, you may end up really begin to think like the people that ran those things and begin to possibly, I'm not saying people are this stupid, but possibly begin to think it was okay and fine and how these systems are run and then become almost defensive about people that talk out against it and say,

Well, no, I think killing people is probably wrong. think, you know, gassing people is probably wrong. think, you know, overdosing on drugs is probably wrong and there's something else to address internally. But having yourself so locked into a system, it becomes who you are. It becomes what you are as well. And if it becomes so embedded into your own personal survival mechanism as a human being, anybody that starts to question your own survival mechanism,

THE HUMAN PROTOCOL (39:55.212)
Yep. Mm-hmm.

THE HUMAN PROTOCOL (40:06.008)
Yep.

James Robson (40:09.601)
you better wait for them to come back at you in some way because woe betide you question their survival mechanism because there is a psychological gap between them thinking there is something wrong with the survival mechanism and the survival mechanism being completely right. There's a psychological gap between that because it's built up and it's protected them, it's protected the individual. If you get locked into that echo chamber, there's a possibility you could end up

THE HUMAN PROTOCOL (40:36.184)
Yeah.

James Robson (40:37.761)
accidentally radicalizing yourself because you think you're defending some ideology and taking it too far to actually perform harm or hurt to another human being. And that's just wrong. That's not.

THE HUMAN PROTOCOL (40:51.918)
And especially, you know, when we become more more dependent on it, that we really are at risk of losing independent thinking, right?

where people are so dependent on these systems and they stop reflecting on us and more kind of directing us without us even noticing. What I've seen from using, course, on our personal lives is that they become, in a way,

guiding you through it, sometimes where you don't want to go next, but little by little telling you and guiding into a direction that that may not be the direction that you want to go. But we become so dependent on it that we stop thinking logically, right? And critical thinking is such a critical aspect of this new future that we just continue to depend on where these LLMs are kind of guiding or pushing us towards.

James Robson (41:51.918)
Yeah, I think that's right. We're heading down a path, unfortunately, where it's the dumbing down of the human psyche because of a reliance on a system that makes our lives a bit more efficient and easier. And we haven't yet solved the bias or hallucinations that pop up through them. And so we're relying on that and then not spending the money on the...

the trained individuals to double check whether or not what's popping out is right or not and you're just moving forward because it makes it easier and so everything ends up taking a shift downwards in the of the level of intelligence and level of kind of human development because of the LLM. Now you know fingers crossed they then you know the LLM will then get smarter because it's trying to get all the people that are at the top of those

sectors and the knowledgeable people and the professors are trying to get that information, that knowledge data in there. That's the hope, but even then how do you trust it without having so... So you know I do have a touch of hope here. I do want to inject a bit of hope which is well isn't it going to mean that we end up getting more quickly to the higher level questions and insights?

THE HUMAN PROTOCOL (42:41.198)
Right.

THE HUMAN PROTOCOL (42:45.858)
That's the hope, right?

James Robson (43:08.995)
because we take away that longer path of all the stuff the LLM is tuned now, great, here it is, like, wow, I got there quick, what's the next level? The LLM can't get me. So you end up almost bypassing all this kind of lower level learning and just get into a better level of learning if you're starting out and then going, now what's next? And so maybe it's a trajectory to getting you there quicker.

THE HUMAN PROTOCOL (43:35.374)
That's the first point, and that's part of what I also see should be some of the positive that AI should bring to us. The concern sometimes is about the journey that it's going to, how we are going to go from this point to that next point that you're telling us, and how that's going to be shaped as these LLMs, they become smarter, right? I also would like to think...

that they become smarter in a way that should drive us into the right direction and not the opposite one. Right? Because some of my concern is when people's life becomes so embedded into that reality, which is talking to these LLMs, it's the perception of reality may be distracted, right? May not be what reality is in a way.

James Robson (44:29.482)
I love what you're saying because things are going off in my head. I'm sorry to speak over you. It's just like, it's all going off because I see this, I'm getting very passionate. I see this balance of, well, who is driving that? Is it driving for consumer-based customer retention, profit improvement pathway?

THE HUMAN PROTOCOL (44:36.3)
It's okay.

James Robson (44:54.207)
legislation in the US around data protection is around consumer and client, or is it being driven for societal benefit and human rights? And then you have you that's going down the societal benefit human rights governance route. And so you end up we end up like circling right background at the beginning of our conversation to go like, which way is it likely to lean? And which one has the, you know,

Is it profit making or is it societal benefit making? Which way is it leaning in the governance?

THE HUMAN PROTOCOL (45:25.238)
And yes, and we have seen it repeating time after time in the era of the social media. Social media was engineered to keep us crawling and on the screen, right?

James Robson (45:38.713)
Yes.

THE HUMAN PROTOCOL (45:39.106)
That's really what social media was engineered for. that's still, even though people say social media is dying now, it's interest media in a way, because you're seeing what you are interested on, whatever. But still, it was engineered to keep people in front of their phones or on the computer, just scrolling, all the attention there. When you don't have no one regulating that aspect, and it's up to these big companies.

James Robson (45:59.023)
and

THE HUMAN PROTOCOL (46:05.654)
right? These AI companies kind of architecting and engineering that part. should make us think on what is the outcomes and what do they want us to get from it and how we're going to get it.

James Robson (46:22.896)
Yeah, you're absolutely right. You've had that recent ruling in the US about Meta and Google actually designing addictive algorithms that affect a person's mental health found against them. You know, there's a great book called Glow Kids, which is a study of how toddlers that have been given maybe a tablet or a phone for many hours at a time, they grow up not understanding the difference between reality

THE HUMAN PROTOCOL (46:30.894)
Mm-hmm. Yep.

James Robson (46:52.85)
and screen image and unreality and so end up with psychological problems very very early on because they're so addicted to the screen time and become very agitated and cry a lot if they have the screen taken away from them. So no, you're absolutely right.

unregulated unfettered access to data is definitely going to help you with the bottom line if you want to do that quickly which is why there probably hasn't been a push for that kind of stuff especially in our age of AI even though I think we're still in complete infancy in AI but even then you know I hope

there may be a correlation between the ruling against the matter of Google in the recent days and the future prospects of AI regulation over the years because, you know, lots of people aren't willing to say that.

The social media that most people have grown up with, I came into it in my kind of late 20s when it was really picking up, has a major impact on mental health. Are people willing to say that? It's like, well, obviously it does because I know if I go to bed and I'm scrolling, because I'm slightly addicted as well as everyone else that has a smartphone, I wake up the next morning not feeling as good as...

I do if I leave my phone outside the bedroom.

James Robson (48:27.298)
You know, it's like simple things, you know, it is affecting your mental health and the whole kind of move fast, break things mentality has actually broken things on a societal mental level and should be changed and regulated early. And so there's more testing. There's more, you know, we've got many years of data now. Let's go back to the data. Let's do the research on the data. Let's centralize that and take a look at what the outcomes are, you know.

THE HUMAN PROTOCOL (48:30.702)
Yeah.

THE HUMAN PROTOCOL (48:56.25)
It's fascinating, all this conversation. We have been talking about the future since we started this conversation, right? And especially when we talk about data as currency, and really at the end of the day, it's really what we're going to be trusting, right? it's trust. It's going to be critical in all this. Without trust, nothing works, right? No society works. No trust in the system. Nothing works. So let's try to look a little bit forward, if we can, even though we

have been doing this from the very beginning. But if we get this right, let's try to dream a little bit more about the future. If we get it right, what does the future look like from your point of view? Who will become this trusted institution in this new world?

James Robson (49:48.337)
Nice. So I think it's going to be a set of organizations that are basically known as trusted data intermediaries. And they are the scary, courageous, transparent organizations that are overseeing technical regulation and technical management to ensure privacy is preserved and ethics also preserved. So you're building ethics into that.

Organisational protocols are in place. People are trained in the right way. They understand how to look at the data correctly. They are thinking ethically around the uses of data.

And then there's contractual, know, do the contracts stack up in the first place because it's great to have this stuff on paper, but then behind the paper, what is the reality of an organization? But actually, if you get the paper wrong, then it really is a flag into, well, if you're to get your paperwork wrong, then maybe you've got your technicals wrong and I shouldn't trust you.

So I think there needs to be a monumental step up of oversight for helping people use data in a better way. So yet again, you know, of course I'm advocating for the field that I work in because it'll mean I'll have a job forever, but I think I will anyway, no matter what, no matter what I say. But there has to be these intermediary organizations that are doing this oversight, but from an ethical point of view.

THE HUMAN PROTOCOL (51:12.588)
All right.

James Robson (51:24.496)
And then even you can do it from the trust scoring that marketeers really love to do around how you feel about an organization. That will very much come down to what they think about how you use that data and their experience of that in the first place. Right through to customer service experience, to data protection experience. Maybe there's an experiential score that...

THE HUMAN PROTOCOL (51:51.19)
Yeah.

James Robson (51:51.33)
and matrices that we end up seeing as some sort of nutrition label on all organizations, you know?

THE HUMAN PROTOCOL (51:58.614)
Yeah, I'm totally on board with everything that you're saying. And that's also how I see it. like, really into this future, the organizations that will win may not be the smartest ones, or those that have the best product or services will be the ones that people trust at the end of the day. As we get more and more into this new future with AI, trust will kind of become

the infrastructure of all these, right? You will go home whomever you trust at the end of the day. Yeah, totally agree with you on that, James. And James, we are approaching the final segment of our conversation, believe it or not. It passed by really quick. So it's kind of one question that I'm going to ask you intended for different groups, okay?

So I'm gonna start with the first group, which is for individuals trying to navigate AI and make sense and use data responsibly. What would be your advice to those?

James Robson (53:03.952)
It's a really good question, it's still a really tough question. I'd say engage a professional governance consultant if you want to try and understand where is the data going? How is it being used? Outline why it's being used in that way? Who has access to that data?

Is the organization, the AI organization transparent with their regulatory framework? Can you control it?

Do you, one thing to really, really watch out for, which I would throw in there, is do you know how decisions were made based on that data? And especially if it's personal data, because if you don't, then you do open yourself up to lawsuits and investigations, because you need to know how you got to a decision if you are asking it to make decisions for you, if it's about other people's lives. I mean, you know.

there's too much hallucination, there's too much proof out there that they're gonna get it wrong. They may find out that I'm actually a Scottish fellow that lives in the Outer Hebrides when I'm actually a Welsh guy that came from Cardiff and lives in London, you know, and it may just get that wrong.

And then if that gets into a court judgment or summons and it's based on this information, you'll just end up with a bunch of false positives around people going through the judicial system and it all crumbles. I've gone dystopian very quickly, but you've got to be really, really careful about the decisions you make based on the data that has come out of a chat bot or an LLM, even though it looks like it's really confident and strong.

James Robson (54:50.065)
strong and wrong at the moment and we're not confident that I think generally that the data that does come out of an LLM is strong enough to rely on it 100%. Do not rely on it 100 % right now.

THE HUMAN PROTOCOL (55:03.916)
Got it. Thanks for that. The next question is your piece of advice for these leaders running these AI companies, right, offering this product and designing this ecosystem. What would be your piece of advice to them?

James Robson (55:21.253)
build protection mechanisms in an ethical way that can be managed by the user. So

Again, it's a very tough question. I got to sit with Sam Altman and say to him, well, what do I think that he should do? I think there should be an ethical button somewhere. Actually be more ethical and then have a feed somewhere saying this is how this was more ethical. Would you like it to ratchet up or ratchet down? And then you can see the comparison between the ethics that come out. And you know,

THE HUMAN PROTOCOL (55:39.662)
Mm-hmm.

THE HUMAN PROTOCOL (55:43.96)
Yeah.

James Robson (56:03.953)
Again, coming back to my next answer, which I know there's a lot of study going into this is what is the level of trust score up against this data and percentage out of 100 of scoring make that far, far more easy to get access to and make that as reliable as you can as well? Is this a 90, 90.8 or is this a 56 because you weren't able to cite?

doctoral thesis that have been a double blind tested for example. Work out what that means to hit a higher trust score and have that as policy. Ethical trust score policy please build that in. Yeah that's what I'd say.

THE HUMAN PROTOCOL (56:46.382)
And transparency at the end of the day, right? I think we all deserve that. If they're using our data, we should have a say or at least access to information on how our data is used and how those systems are architected. yeah.

James Robson (57:02.929)
Yeah, yeah, yeah, I definitely want to come on to that one. I mean, we're getting back into the, well, if your information is like money and it's being used in that way, wouldn't you like to know where it is, who's using it, why they're using it, who's accessing it? And being able to manage that. Should be a no-brainer. Build that one in, 100%, 100%.

THE HUMAN PROTOCOL (57:23.47)
It should be a no-brainer, right?

Totally agree. And last but not least, James, to governments regulating AI, what would be your piece of advice there?

James Robson (57:40.467)
Wow, you hit me with the small questions. Yeah, no, that's super tough. Now, what I am really liking is the fact that the regulator in the UK has set up a sandbox facility or is in the process of setting up a sandbox facility to be able to test mechanisms of use of data that may actually sit outside the law.

THE HUMAN PROTOCOL (57:45.282)
Sorry.

James Robson (58:08.634)
because sometimes you may need to work out how to align to the law, but you can't work out how to do that without having to step over that line in the first place. So it's set up this regulatory sandbox, or is about to, it's working out how to, it's gonna spend the next year working out how to build a sandbox where it could potentially break the law. They'll come back to me and say like, no, we're not doing that, potentially break the law here.

very, very contained and then work out how to align to the law. Then you may go, well, the law doesn't meet the level of requirement for this use case. And actually, we don't think this is ethically wrong. Why not build those mechanisms into legislation to allow these incredible sandbox environments that are so highly secure and contained?

but tested and researched and double blind tested again to go cool, now let's meet regulation at the speed of this testing to step up along with it because there's the argument that as soon as legislation is passed, it's dead because technology has moved 10 steps forward. Why can't you build that alignment with a sandbox to level up?

on a regular basis with the technology so you have a consistent running sandbox that's able to meet the speed of regulation needing to keep up. That's what I'd say.

THE HUMAN PROTOCOL (59:36.862)
It's going to be fascinating to see what's happening in the next couple of years, because all this is accelerating at the speed of light. In a way, right? So I'll be pretty much looking forward to see what happens next and what we see from different governments around regulation. So yeah, this conversation wasn't just about data and AI, it was about trust at the end of the day. Trust is really the foundation of everything we're building next, right?

James Robson (59:44.742)
Yeah.

THE HUMAN PROTOCOL (01:00:05.792)
And I want to take a moment again, James, to appreciate you one more time for being with us today.

James Robson (01:00:12.838)
Yeah, you too. Thank you so much for having me on. Appreciate that.

THE HUMAN PROTOCOL (01:00:15.618)
Perfect, it was our pleasure and if this episode made you think differently about data, share it with someone building AI, leading a company, or even shaping policy. Follow the human protocol because the code may change, but our humanity is the constant. Thank you all, talk to you next week.