OpenAI funds research on ‘moral AI’; to understand

Developer 1701
OpenAI financia pesquisas sobre ‘IA moral’;   entenderAll about Artificial Intelligence see more All about OpenAI see more OpenAI Inc., the nonprofit that develops ChatGPT, has awarded a research grant to professors at Duke University for a project called "AI Behavior Research."

The grant is part of a larger three-year program worth US$1 million (about US$5.8 million) and will help investigate how artificial intelligence can be used to make ethical decisions in healthcare law and business .

Participants in OpenAI-funded research have already investigated how AI can be a 'behavioral GPS'. Walter Sinnott-Armstrong, professor of behavior at Duke, is primarily responsible for the project, along with Jana Borg. The two have already done research into how AI can act as a “moral GPS,” helping people make more ethical decisions, like deciding whether to receive a kidney.

The problem faced by researchers supported by OpenAI is technology, ethics and philosophy (Image: Svet Photo/Shutterstock) Not much information is known about this OpenAI research. But according to a press release, the focus of the research is to teach algorithms to predict human behavior in conflict situations.

This is difficult for AI. This is because machine learning, like that created by OpenAI, is basically a statistical machine that learns patterns using large amounts of internet data.

Read more:

OpenAI made a big mistake - and it could cost them What is Sora, OpenAI's video-making artificial intelligence? Q*: What do we know about OpenAI artificial intelligence? These examples may predict the next word in a sentence, but that doesn't mean they understand things like behavior or empathy. Morality is a matter of fact. It changes depending on culture, situation and experience. And there is no answer on that.

Therefore, the challenge for researchers at OpenAI and Duke is not just technological, but also ethical and philosophical. How can you teach a machine something that humans can't explain in the same way? By 2025, when the subsidy ends, we may have a solution. Or more questions.

Apple plans to make Siri similar to ChatGPT and Gemini Live Apple plans to make Siri a virtual assistant similar to OpenAI's ChatGPT and Google's Gemini Live. "LLM Siri" (Siri language model) allows users to have natural interaction and conversation with Siri.

Apple plans to update Siri, which could be more similar to ChatGPT (Image: Apple/YouTube) One of Apple's plans to integrate AI into the iPhone is to improve Siri's ability to perform advanced tasks. To achieve this, the assistant would have “expanded” the ability to work with other candidates.  And I use Apple Intelligence to summarize and write notes.

Find out more about Apple's plans for Siri in this article from Olhar Digital.