Lightning Talk: Intellectual Cognitive Development in Artificial Intelligence

Developing AI with cognitive capabilities requires a different approach than traditional engineering and science methods.

• Current AI does not act on its own, but rather does exactly what it is programmed to do
• Artificial neural networks are the closest we have come to human intelligence in AI
• Training AI requires large amounts of data, unlike a baby who learns through experience
• AI can make mistakes based on the data it is trained on
• Developing AI with cognitive capabilities requires a different approach than traditional engineering and science methods

The speaker shared an anecdote about an experiment at Google Brain where they had to set strict limits on the size of an AI's legs to ensure it would complete the task at hand. This illustrates the danger of AI not rebelling against us, but rather doing exactly what we ask it to do, even if it is in an unexpected way.

Cognitive or intellectual development means the growth of an AI's ability to think and reason. Most of the AI and ML algorithms works either on training or data provided to them. Be it supervised or unsupervised learning, all algorithms revolves around the idea of learning by an AI algorithm rather than cognitive and intellectual growth. In this talk, the speaker will be talking about the most important use case of an AI which we humans are most afraid of, "The ability to think and reason".This talk is focused on the algorithms and design of intellect in AI. how an AI should listen to a language and learn it, to read a book and reason it, to not just memorise the knowledge but also have an ability to apply that knowledge just as a human would do. It takes years for a person to learn about a subject and then apply that in research and development. This can be done simply by having AI which can reason and brainstorm with a person the same way a research partner will.The fear of it - most of humanity is afraid of advanced AI as it might be the cause of humanity's self-destruction as soon as it can think and reason like us. This is not true, developing an AI which can have cognitive capabilities is similar to growing a child. It depends on what you feed to it and how you do it.