OpenAI Competitor: Its Chatbot has a Rudimentary Consciousness
AI chatbots are notorious for inventing things and spewing out bigoted garbage. A firm founded by former OpenAI researchers is taking a different approach: teaching AI to be conscientious.
Wired reported that Anthropic, an OpenAI competitor, has built a chatbot called Claude with a set of rules, or ‘constitution,’ that is based on the Universal Declaration of Human Rights, and other sources, to ensure the bot is powerful but also ethical.
Jared Kaplan is a former OpenAI researcher consultant, who founded Anthropic in collaboration with a group his former colleagues. He told Wired, Claude, in essence, learns right from wrong, because the training protocols \”basically reinforce the behaviors which are more in accordance with the constitution and discourage behaviors that can be problematic.\”
Source:
https://futurism.com/the-byte/openai-competitor-chatbot-ethics
