I like to experiment. I view myself somewhat of a marketing engineer, constantly looking at ways to break down tools, methods, processes, and technology, to find out how it all works, before I put it back together again. I did this when I was a kid with radios, model train sets, and pretty much anything I could get my hands on.
As an adult, this has shifted into my work in marketing and branding. This one particular project, the TruePoint AI Data Model is something that others might also find interesting. Let me tell you a little bit about it and see what you think.
Let me start at the beginning …
I am a news-o-file, if that’s a thing. We used to be called Political Wonks. Ever since I remember, I’ve been interested in politics. I got my undergraduate degree in Political Science, dabbled in a variety of campaigns, and have pretty much been following Politics as if it were a sport. I learned the theoretical aspects of government, policies, diplomacy, and a bit of economics in college. But it’s only after being in the trenches that you realize politics is not about policy. It’s something else. Something performative.
I’m not super convinced we’re not a simulated cable TV show in the future poking fun at a disfunctional global political landscape. I say that without bias or political party affiliation. Nobody takes it seriously anymore, other than to feign anger and outrage when one side does something mildly discomforting.
Now, where was I?
When I started the TruePoint Project, I wanted to uncover hidden biases in news, social media, and reporting. Not the obvious biases, but the subtle linguistic clues, the small choices in wording and phrasing that shape our perceptions quietly yet profoundly.
I began gathering massive amounts of data, experimenting with algorithms to identify those subtle linguistic patterns. My aim was simple but ambitious: to create a neutral, reliable way to measure factual accuracy and source credibility.
The beta version of the application was able to take an online news article, run it through our algorithm and data set to look for bias and truth markers. It would then tabulate a score on a spectrum of neutrality. The idea is that the more neutral an article may seem, the more likely it is to be balanced, and therefore contain the best opportunity to be true. That was the hypothesis, at least. The system works. Really well, actually. I am able to determine bias for articles, authors, sources, media networks, and it’s now grinding through social media profiles.
The project took an interesting turn as I began experimenting with how Artificial Intelligence language models are trained. With AI-generated content becoming common everywhere, I found myself asking: Could my model distinguish genuinely human writing from AI-generated text? Could I teach an algorithm to recognize the unique signatures of authentic human expression?
After a lot of research looking into various LLM’s, their training methods, and potential to bias the data set over time (more on hacking the AI systems in a future lecture), the answer was clear. I discovered a series of distinct linguistic and stylistic markers—unique fingerprints—that AI hasn't yet mastered replicating. TruePoint AI can now reliably detect AI-generated content.
But even more exciting is that it can also help train AI to capture those unique human voices, tones, and styles more effectively. Since I know we are not going away from machine generated content, we can at least give it a little personality.
That’s where I brought in my brand foundations methodologies for developing brand voice and tone system. I have been doing a lot of brand consulting lately and this is something I’ve developed a pretty effective process for. The results is that I can now generate a unique character persona or archetype for content creators using AI language models. It’s a way for organizations to produce high quality content, that maintains a genuinely authentic and human voice. The final goal isn’t necessarily to be able to generate all content free of humans, rather the opposite. It’s being designed as a tool that copywriters and content creators can use to gather research, vet ideas, get reliable quotes/sources, and publish content.
For marketers, brands, and creators, this means maintaining authenticity and genuine human connection while embracing powerful technology.
TruePoint isn't just about identifying bias or authenticity; it's part of a bigger mission—exploring how technology shapes the future of marketing and branding. Ultimately, my hope is to ensure technology enhances, rather than replaces, our human ability to communicate authentically.
Forseti is the ultimate icon for finding the truth because he represents justice without bias, wisdom without arrogance, and resolution without violence. Unlike many Norse gods who rule through might and conquest, Forseti wields the power of reason. He is not a warrior, but a judge—his strength lies in his ability to cut through deception, mediate disputes, and guide even the most stubborn minds toward fairness and truth.
His hall, Glitnir, is a place of radiant clarity, where all who seek resolution come to be heard. The very walls shimmer with silver and gold, symbolizing transparency and enlightenment. In Forseti’s presence, arguments are stripped of manipulation, half-truths dissolve, and only the unshakable foundation of honesty remains.
In a world full of marketing fluff, corporate spin, and social media distortions, Forseti is a perfect symbol for cutting through the noise and finding what’s real. He doesn’t play favorites, doesn’t embellish, and doesn’t tolerate deception—only the pure, unfiltered truth stands in his domain.
