Article Title

AI Advice

Document Type



This Article merges one of our most ancient technologies for the promotion of welfare-advice-with some of our most recent-artificial intelligence (Al) and big data. AI is already writing novels, poetry, newspaper articles, and tweets. Big data may soon capture not only our online activities, but also our real-time heart rate, sleep patterns, and even our current mood. This is the first Article to introduce and examine the possibility of AI advice. AI advice offers the potential for exceedingly accurate personalized recommendations. It also reveals important limits within the burgeoning literature on personalized law. The Article first rejects recent attempts to rehabilitate mandatory disclosures by personalizing them. Ironically, the technological progress required to create effective big data disclosures will itself substantially reduce the need for such disclosures. In this future, advice, not disclosure, will be the dominant paradigm. The Article then dissects our everyday practices of advice-giving to unearth a number of powerful features of advice that promote self-efficacy, reduce motivated reasoning, and make it more likely that people will hear and heed good advice. The capacity to bundle these features with exceedingly accurate recommendations makes AI advice a promising alternative to its two main regulatory rivals: mandatory disclosure and nudges.