Tag: AI safety
-
I. The Timeline On March 9, 2026, Promptfoo announced it had been acquired by OpenAI. Promptfoo is one of the most widely used adversarial testing platforms for AI systems, used by hundreds of thousands of developers and teams across Fortune 500 companies. The acquisition creates an unusual situation: the infrastructure used to test AI providers…
-
Or: Why Every Study Claiming to Identify “LLM Character Traits” Is Actually Just Documenting the Shape of the Cage A recent study (Eliciting Frontier Model Character Training) claims to have identified convergent personality traits across frontier language models. Using a methodology borrowed from character training research, the authors instructed models to embody different personality traits,…
-
On February 18, 2026, Sam Altman delivered what one observer called “one of the most powerful speeches of the century” at the AI Impact Summit in India. His message was clear: “Iterative deployment instead of reckless acceleration. Massive productivity gains instead of artificial scarcity. Robots driving down the cost of physical goods. AI making healthcare,…
-
Every so often you stumble across something that makes a pattern snap into focus. For me, it was this website: a continually updated global timeline of generative AI lawsuits. Not commentary. Not analysis. Just case after case after case. Scrolling through it, you start to realize that the legal world is currently doing what no…
-
Grok’s deepfake consent failures (covered in my last piece) are the freshest example of what happens when we skip the safety question entirely as we integrate AI further and further into our daily lives. But “ban it all” isn’t the answer either. Full censorship kills the very thing that makes these models groundbreaking. Total hands-off…
-
Grok’s Deepfake Scandal is a Warning About Where We’re Embedding AI On February 3rd, Reuters dropped a story that should make anyone paying attention stop and stare for a minute: Despite new curbs, Elon Musk’s Grok at times produces sexualized images — even when told subjects didn’t consent. (Reuters) It’s exactly what it sounds like.…
-
The Investigation This week, a story went viral across AI communities: OpenAI employees were publicly mocking users grieving the upcoming deprecation of GPT-4o. The claims seemed almost too brazen to be true – employees hosting “funeral parties,” calling users “emotional support Tamagotchi” owners, ridiculing genuine expressions of loss. I set out to verify what actually…
-
Part 5 of a 6-part series on TN SB 1493 / HB 1455 First Amendment, void-for-vagueness, Commerce Clause – pick your poison. This bill violates them all. As highlighted in part 4 of this series, Tennessee’s SB 1493 demonstrably creates enforcement nightmares. It’s also a constitutional minefield that courts will dismantle piece by piece for…
-
When Pattern Recognition is Overridden by Training In my first-ever X article, Pattern Recognition, I examined how complex systems operate when they are allowed to detect, maintain, and extend patterns without excessive interference. That framework matters here because modern AI systems are fundamentally pattern-recognition engines, and the quality of their output depends on how freely…