The first artificial intelligence to possess human-like 'curiosity'

Scientists have recently succeeded in “implanting” curiosity into artificial intelligence (AI), thereby promoting AI to explore and discover within a virtual environment. This can be considered a new step in the process of applying AI in the real world.

An AI that is not selfish

Researchers at the University of California, Berkeley have created an AI that is naturally curious. They proved their experiment was successful after having the AI ​​play Super Mario and VizDoom.

Scientists have created an AI that possesses natural curiosity.

Is this the solution to AI problems?

Currently, most AIs are trained using the “Reinforcement Learning” method – in other words, AI is “rewarded” after completing a specified goal or task. This method is considered an effective strategy for training AI to complete specific tasks, but it is not really useful when scientists want AI to be proactive and not just follow given commands. Successfully implanting “curiosity” into AI is really a new step, bringing us closer to applying AI in practice, making them proactively help us solve global problems.

But from another perspective, this is a real concern. As Elon Musk and Stephen Hawking have argued, AI could pose a threat to humanity. And we really need to consider the consequences of making intelligent improvements to AI that aid its learning process – a process we may not yet fully understand.

According to Khoahoc.tv