Anyone who has seen the movie Matrix would remember feeling the chill of a threat going down their spines when the extent of AI control over the human world was revealed. Ever since machinery has become a part of the mainstream production, the concept of machines being a threat to humanity has compounded. After the emergence of AI, this fear has only escalated further because of which AI is being presented as a serious threat to humanity.
The ones that are stressing about AI becoming a threat to humanity are not just normal people, unaware of the intricacies of AI technology but are eminent scientists and experts. Elon Musk, Bill gates are just a few of the names who have been vocal about the imminent threat of AI overtaking the world and seem to be concerned about the growing powers of AI. Is this worrying really valid? Before answering that question, it becomes important to understand how AI will become a threat to humanity.
How AI poses a threat to humanity?
One argument states that as humans are the dominant species because of their intelligence; if AI achieves human intelligence or surpasses it, then it will become too powerful and would be difficult to keep under our control. Once AI learns how to acquire knowledge it would keep acquiring it at an unprecedented pace which cannot be achieved by humans.
The other concern is that AI would be able to acquire the human thinking capacity but programming human values into it would not be possible and thus it would behave very differently than humans.
Will AI become a threat to humanity?
AI would not overpower humanity because of a ‘desire’ for power but is said that AI could overpower humanity because of a lack of clarity of goals. Let us consider the favorite example of Bostromm, a Swedish philosopher. He asks us to imagine a machine which is programmed to make paper clips. Over the passage of time, it will keep getting smarter and more powerful. Eventually, it will develop other machines for creating paperclips but as it is super intelligent, humans would no longer have the ability of programming it or over-riding its programming. It would never develop human values and thus would not care about the consequences of its actions. It would eventually begin to convert everything into paper clips. Maybe even humans.
How real is the threat?
AI would become a threat to humanity when a computer program is able to evolve itself. But what this theory forgets to mention is the fact that imparting the cognitive ability to machines isn’t something that we have achieved and it isn’t something we are working towards. The area of AI research and applications deals with processing data and making predictions based on it. It is not about giving machines or programs the ability to imitate the entire functioning of the human brain, more so when we aren’t clear about how the human brain works.
Considering the paper clip machine scenario, it does not seem likely that humans will configure a machine which is intended for the purpose of creating paper clips to be so advanced and sophisticated that it would be able to overtake humanity. Moreover, such a machine is not likely to be fitted with weaponry or have access to programming which could let it control it and would thus not be life-threatening to humans and it would be possible for humans to manually shut it if it gets out of hand.
What can we do to lessen the threat?
As is apparent now, we need to practice responsible and ethical practices when it comes to dealing with AI. It would take just one irresponsible human act, when AI technology is more mainstream, to create a machine which would become super intelligent or what’s worse is it could reach in the wrong hands, where this power will be misused.
The only solution is to restrict the use of this technology in a safe environment. AI could turn into a potential threat but the progress that could be made by AI could also not be ignored or given up, therefore the onus falls on us to make responsible choices about what we do with this technology.
Thus, it could be said that AI won’t be a direct threat to humanity. It could, however, cause widespread damage to human lives, as a by-product of achieving some other goal. Responsible practices and the right regulations and legal framework are needed to make sure this technology doesn’t fall into the wrong hands.