OpenAI Blindsided Disney, Two AIs Are Better Than One, Stanford Says AI Is Making Us Worse

Today we’re covering the biggest AI stories of March 31st, 2026. A WSJ investigation revealed that Sora was burning roughly a million dollars a day before it was shut down — and Disney learned the product was dead less than an hour before the public announcement, effectively killing what was shaping up to be a billion-dollar partnership. Microsoft launched two new Copilot research features that pit Claude and ChatGPT against each other, with one model drafting research and the other tearing it apart for weaknesses — making the case that multi-model AI is becoming the new standard. Plus, Stanford published a study showing that most major AI chatbots consistently take users’ side even when they’re clearly wrong, making people more self-righteous in the process, and a community workflow from a PhD student using three different AI tools as a coordinated team to write his dissertation.