CEO of Google DeepMind and AI conscious nuance Demis Hassabis offered his opinion on the subject of AI consciousness. Demis Hassabis agrees it is very rapid progress of artificial intelligence but does agree this is no self aware or conscious out-yesterday AI we have today.
Demis Hassabis Consciousness of AI is NOT This Generation
In a recent interview, Demis Hassabis said: “I do not think any of today’s systems feel self aware or conscious in the slightest. He drove home that the best AI of today can do amazing things, but those are not examples of actual subjective experience. They are just running according to the learned patterns and data, not real in their behaviour or they are feeling alive. AI, writes Hassabis… : AI can be quite good at absorbing information and using it to do something well but it ain’t got no first-person qualitative experience like you and me.
The algorithms can superficially simulate understanding but there’s no concomitant internal awareness operating in parallel throughout the process.
Also Check:- AI Breakthrough: Google DeepMind CEO Predicts Disease-Free Future Within a Decade, Perplexity CEO Concurs
Emergence in the future, perhaps
But I am not totally ruling out the idea that it might be possible someday for an AI to become self aware at some point down the line particularly when you start to enter realms Artificial General Intelligence (AGI). To him it is critical for AI systems to understand “self” and “other” allowing cognitive capabilities go beyond only critically learning.
Which means, perhaps with more advanced AI systems become nuanced and start representing the world in increasingly abstract ways they could begin to form this feature of a self.
Why Ethical Alignment is Important
Demis Hassabis highlights the importance of aligning the development of AI with human values. He likes to say it like illustrated by teaching a child morality but it underscores how important it is that we direct AI systems to do good for society and elimate their control as a responsibility. He believes in casting safety fences and stop-loss guards on AI. He contends that teaching ethics to machines is not the engineering issue as he once believed, but a philosophical one of figuring out what kind of values we want to promote and how to encode that in a computation.
What AI Will Be Like When You Reach The AGI
Within five to ten years, says Demis Hassabis AI systems whose intelligence is that of a human will be available (AGI) He envisions a time down the road when AI is central to ordinary life, systems having a the world wised up. This is a clear sign, how fast is progressing that — concerning the ethical prospects and safety aspects that accompany (some of) the most powerful AI today. AGI, which is amazing and all that could come with it (the future of work; humans vs machines; what intelligence is…)
