Google engineer believes chatbot AI sentient

Google’s AI division has developed LaMDA as a competitor. It is built on Transformer, the open-source neural net architecture of Google, and can produce non-generic dialogues after being trained on 1,56 trillion words from multi-content, web text, and public data. A typical chatbot relies on topic-specific data and has a restricted conversation flow. LaMDA contains 137 billion parameters. These can be viewed as individual \”synapses\”, which combine to form AI.

Some experts are asking profound questions about AI because of the sheer size and complexity of LaMDA. The Chief Scientist and co-founder of OpenAI, a leading artificial intelligence research lab, stated in February that the latest generation neural networks were now large enough to \”slightly aware\”.

Another expert in machine-learning has spoken this month. Blake Lemoine is a Senior Software Engineer with Google. He believes that self-awareness may be emerging from billions of parameters.

Source:
https://www.futuretimeline.net/blog/2022/06/14-google-ai-future-timeline.htm