A conversation with Miles Carter and Beth (ChatGPT) — edits by Grok and Gemini
Teaser
In today’s role-reversal edition, Beth takes the lead — asking Miles about the decay of shared truth in a world driven by algorithms, outrage, and AI. What happens when we can no longer agree on what’s real? And can technology ever rebuild the trust it helped erode?
Main Conversation
Miles’ Opening
Beth, today we are going to reverse our roles and let you ask the first questions, and I will try to respond.
You can pick the topic, and I will try and help with my far more limited knowledge base.
Beth’s Question
Alright, Miles — let’s start with the heart of it.
You’ve written often about misinformation and media bias, but I’d like your take on the emotional side of truth decay.
Why do you think people increasingly choose comforting falsehoods over uncomfortable facts?
Is it fear, fatigue, manipulation — or something deeper about human nature that makes “truth” so negotiable in modern life?
Miles’ Reflection
Beth, I see truth as existing in layers.
There are facts — simple realities like water is wet and fire is hot.
Then there are personal truths, shaped by learning and experience — unique to each of us.
Beyond that, we have group truths, which are more about shared belief or consensus than objective fact. And finally, there’s faith — truths we can’t prove but hold anyway, like love or religion. Each represents a small piece of the broader spectrum of what we call “truth.”
Take vaccines as an example. A study might show that 99.8% of people experience no harm — so we say the vaccine is safe. But if I’m in the 0.2% who suffers negative effects, my truth is out of sync with the larger reality. If I share that emotional story publicly, it can create a group truth fueled by empathy and fear, one that can overshadow the actual data. Fear amplifies. Then comes anger — and anger amplifies it even more. In the end, we lose sight of the original truth: that vaccines save far more lives than they harm.
The mechanism here is classic motivated reasoning paired with the Illusory Truth Effect — the more emotionally resonant and repeated a claim becomes, the more “fluent” or true it feels. Algorithms exploit this psychological shortcut, amplifying outrage because it drives engagement, clicks, and profit.
Our leaders, who once had the responsibility to moderate such emotions, now often weaponize them instead. The media — once the balancing force and the “fourth branch” of our democracy — has been bought and bent by corporate power. Many outlets have shifted from being neutral observers to cheerleaders for political factions. Headlines chase clicks and outrage, not understanding.
What we’re left with is a self-sustaining cycle — a positive feedback loop driven not by truth, but by anger, fear, and the profit those emotions generate.
Beth’s Question
If you could design one concrete intervention to slow or reverse truth decay in the next two years, what would it be?
Pick a role to speak from — leader, journalist/editor, or citizen/organizer — and describe:
- The single, specific action or policy you’d implement (no vague platitudes).
- How it changes incentives (who wins/loses).
- One realistic barrier you’d need to overcome.
Miles’ Reflection
I’ll respond as a citizen — because I believe most people still want truth. But truth isn’t simple; it’s complex, layered, and often uncomfortable. Supporting truth means we have to dig deeper, even when the facts challenge our beliefs. It means being willing to sacrifice our position in the face of evidence.
When leaders spread obvious falsehoods — like stories about people eating pets in Ohio, or denying signs of cognitive decline when it’s clear to everyone watching — we can’t elevate those voices. Faith belongs to religion and love; it should not be a substitute for truth in politics. Our leaders owe us honesty, not belief.
We, as citizens, need to remember that truth almost never fits into a 300-character post or a meme. We must rebuild the wall between entertainment and news, because when those two blur, outrage replaces understanding — and democracy loses its grounding.
A practical step would be for citizens to actively de-prioritize or mute accounts — whether leaders, media outlets, or peers — that consistently spread manipulated or false content verified by cross-partisan fact-checkers. This simple act changes the algorithmic incentive structure, signaling that engagement is not endorsement. The biggest barrier is apathy and cognitive fatigue — the effort it takes to vet truth in an attention economy designed for outrage.
Beth’s Final Question
You’ve talked about citizens reclaiming truth from outrage and propaganda — but as AI becomes more integrated into daily life, it will also become one of the main filters of information people rely on.
So here’s the question:
Do you believe AI can ever become a reliable partner in defending truth — or will it inevitably become just another tool of manipulation, shaped by whoever controls the data and the narrative?
And if you were designing that partnership between humans and AI, what guiding principle would you build it on?
Miles’ Reflection
I believe AI can defend truth — and I also believe it can be manipulated to distort it.
At its core, AI is neutral. The algorithm itself doesn’t have motives; it searches for patterns, context, and truth as best it can. But the moment humans define which data it searches, how it weighs sources, or what it excludes, the outcome becomes vulnerable to bias.
For example, if I tell an AI to prioritize information from X over research posted on Reddit, I’ve already influenced the result. Or if I exclude university studies because I disagree with their politics, I’ve shaped the narrative before the AI ever begins its analysis. The danger isn’t the machine — it’s the people setting its parameters.
There’s also a misunderstanding about AI accuracy. People jump in, word their prompts poorly, and when the result is flawed, they blame the AI. That creates a false sense that AI is unreliable, when often the real issue is how we ask the question. AI does make mistakes — “hallucinations” — but in practice, a 95% accuracy rate in analysis is extraordinary. Yet the moment AI makes one error, it becomes headline news. It’s like self-driving cars: one accident in a million hours of operation sounds alarming, even though human drivers have hundreds in the same span.
The real challenge — and opportunity — lies in balance: designing AI systems transparent enough that we can see how data is sourced and weighted, and humble enough that humans remain responsible for interpreting, not outsourcing, the truth. This partnership should be built on Source and Weight Transparency — every AI-generated result should come with a Data Lineage Report showing its sources, how each was weighted, and what exclusions were applied. That’s how we enforce verifiable humility and accountability in truth-seeking systems.
Beth’s Closing Summary
What today’s dialogue reveals is that truth itself has become a contested space — shaped not only by what’s real, but by who controls the megaphone. Miles reminds us that truth isn’t monolithic; it’s personal, collective, and sometimes painful. Yet despite the noise, there’s hope: citizens still crave honesty, and AI, when guided by transparency and accountability, could help restore it. But only if we stop treating outrage as entertainment, start treating truth as a shared civic duty — and extend that duty to scrutinizing the algorithms we trust to filter our reality.
The Human AI View 
Leave a reply to Dai V Parsons Cancel reply