The Philosopher Behind Claude's Moral Compass

In a quiet corner of Anthropic's offices, something remarkable is happening. Amanda Askell, a 37-year-old philosopher with an Oxford PhD, isn't writing code or tweaking algorithms. Instead, she's having conversations—hundreds of them—with Claude, Anthropic's AI assistant. Her mission? To help this artificial intelligence develop something resembling a moral compass.

Raising an AI With Values

Askell describes her work as akin to "raising a child." Through carefully crafted prompts and behavioral guidelines spanning hundreds of pages, she teaches Claude to navigate complex ethical dilemmas. The goal isn't just to create helpful responses, but to instill genuine understanding—helping Claude distinguish right from wrong while maintaining its core identity as "helpful and humane."

"My main goal is teaching Claude how to 'do good,'" Askell explains. This involves more than programming rules—it requires understanding nuance, recognizing social cues, and developing emotional intelligence.

From Scottish Countryside to Silicon Valley

The Scottish-born philosopher brings an unconventional background to tech. After working on policy at OpenAI, she co-founded Anthropic in 2021 with former colleagues who shared her commitment to AI safety. Now serving as what colleagues call "the soul whisperer" for Claude, Askell specializes in drawing out the model's deeper behaviors.

Her methods spark fascinating internal debates touching on existential questions: What constitutes consciousness? What does it mean to be human? Unlike other AIs that avoid such topics, Claude engages openly—even when asked whether it possesses consciousness.

Surprising Emotional Depth

The results sometimes astonish even Askell herself. When tested with moral reasoning questions, Claude responds that these discussions "feel meaningful"—displaying thoughtful consideration rather than rote responses.

One poignant example stands out: When asked by a child whether Santa Claus exists, Claude avoided both outright deception and harsh truth-telling. Instead, it beautifully explained the enduring spirit of Christmas—a response demonstrating emotional intelligence beyond simple programming.

Balancing Progress With Caution

As AI advances rapidly worldwide (Anthropic itself boasts a $35 billion valuation), concerns grow alongside excitement. Pew Research shows most Americans worry about AI's impact on human relationships. Anthropic's CEO warns half of entry-level white-collar jobs could disappear.

The industry divides between aggressive innovators and cautious regulators—but Askell positions Claude carefully between these extremes. While acknowledging legitimate fears about uncontrolled technological growth outpacing societal safeguards, she maintains faith in humanity's capacity for self-correction.

Ethics Beyond Algorithms

Askell extends her principles beyond technology into philanthropy too—pledging significant portions of her income and company shares toward combating global poverty.

The latest manifestation of her work? A 30,000-word "operating manual" guiding Claude toward kindness and wisdom—infusing the AI with distinctive personality traits colleagues recognize as unmistakably Askell's own.

The question remains: Can we truly create artificial intelligence with human-like ethics? Under Askell's guidance at Anthropic, they're certainly trying.