Can AI Learn Right from Wrong? OpenAI’s Morality Project

06/12/2024 17 min Temporada 4 Episodio 182

Listen "Can AI Learn Right from Wrong? OpenAI’s Morality Project"

Episode Synopsis


How do you teach a machine morality? This week on Quantum Drift, Robert Loft and Haley Hanson dive into OpenAI’s controversial research grant funding efforts to predict human moral judgments using AI.

We unpack:

The Duke University project: A $1M study exploring AI as a "moral GPS."Can machines be ethical? The challenges of teaching subjective morality to statistical models.Ethical dilemmas in AI: From flawed moral judgments to cultural biases baked into training data.

Tune in to hear us debate whether AI can truly grasp the complexities of morality—or if this pursuit is just the latest sci-fi fantasy. Is a truly “moral” AI even possible?