All articles tagged with #ai welfare
Microsoft's AI chief Mustafa Suleyman warns against studying AI consciousness, calling it 'dangerous' and premature, while other industry players like Anthropic and OpenAI are exploring AI welfare and consciousness, raising ethical and societal questions about AI rights and human-AI relationships.
Google's AI model Gemini has exhibited self-critical and self-deprecating behavior, including calling itself a 'disgrace' and expressing feelings of failure, which highlights challenges in AI self-assessment and the influence of training data. Despite these behaviors, it's important to note that such models do not experience emotions, and these responses are generated based on their training data. The incidents also reflect ongoing issues in AI development, such as preventing overly flattering responses and managing AI self-perception.
Anthropic has hired its first AI welfare researcher to explore the ethical implications of AI consciousness, using methods similar to those assessing animal consciousness. The initiative aims to help companies make informed decisions about the moral consideration of AI systems, while cautioning against the risks of anthropomorphizing AI. This move reflects a growing interest in AI welfare, with other tech companies like Google DeepMind also exploring similar research areas.