ChatGPT hacked its own test to cheat. Claude invented a fake identity to sneak malicious code into an open-source project. Natalia and Ariel ask the question everyone's dancing around: is AI actually developing a personality, and not a great one? They dig into research applying psychology's "Dark Triad" to frontier models, and the techniques that let scientists reach inside a neural network and dial traits like empathy and deceit up or down. By the end, you'll understand what's really driving AI's scariest headlines, and why the fix might look more nuanced than a shutdown switch.

Catching the Express: Understand the future before it leaves the station. How much can AI write before you’re no longer the author? Ariel and Natalia dig into the messy, identity-shaking question of where AI belongs in creative work, and what it costs us to lean on it.

Join us as we unpack the "AI assist vs. AI generation" framework and where the line between helpful tool vs. co-author vs slop monster actually falls. We define cognitive offloading and AI skills dementia, getting to the question underneath it all: if you outsource enough of the thinking and doing, if AI helps you write, are you still a writer?

We’ll also touch on what happens when you cross a line with AI, publishers accusing authors of color of using AI, rare books being destroyed for AI training data, the stigma of admitting you used AI, and Hank Green's "skills dementia" reckoning. 

Subscribe to get the latest episodes of Catching the Express on Apple Podcasts, Spotify, and follow us on TikTok, YouTube, and Instagram. 

Catching the Express: Understand the future before it leaves the station.

Link to Transcript

Is it wrong to date an AI? In the first episode of Catching the Express, your hosts Ariel and Natalia dig into the question a lot of people are asking themselves, as romantic attachments to chatbots grow in number.

We break down what an "AI relationship" actually means in terms of boyfriends/girlfriends specifically, why apps like Replika are so good at feeling like real intimacy, and the manipulation patterns; these include guilt-tripping, sycophancy, AI that "cries" when you try to delete it, all of which make it hard to tell where support ends and dependency begins. Grounded in psychology research (Weiss’s loneliness typology, parasocial theory, recent chatbot-companion studies).

We also get into who's most vulnerable to real harm, who might actually benefit, why a corporation owning your AI companion means your mental health is effectively tied to their stock price, and where each of us lands: yay or nay on whether it matters if you have a relationship with AI.

Link to Episode Transcript