Artificial Intelligence – Promising or Catastrophic
Published on : Monday 30-11--0001
How often have you heard these two words from someone? It definitely seems like the talk of the town right now. So what is AI really? The term artificial intelligence was coined in 1956, but AI has become more popular today thanks to increased data volumes, advanced algorithms, and improvements in computing power and storage. Artificial intelligence is nothing but the theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision making, and translation between languages.
Did you just hear translation between languages? Yes, that’s right. The Google translator you’ve been using to keep up with your French lessons is actually based on the artificial intelligence that you’ve been using unknowingly.
AI automates repetitive learning and discovery through previously recorded data. AI performs frequent, high-volume, computerised tasks reliably, and without fatigue. However, for this to work, human inquiry is still essential and it is important that the humans ask the right queries. This basically leads to storing of cookies and search history. In layman’s terms cookies will store our previous searches and these searches will tell the webpage the preferences of the user, so the next time you log in and search for something, you are more likely to see something relevant to your interest.
AI is not a single product sold in the market on its own. It is packaged with other existing products to make it of more useful to the consumer, a very basic example being the existence of Siri on an iPhone or an application of Google maps on our android phones. Progressive learning algorithms are the fulcrum of AI. It finds structure and regularities in data so that it can acquire a skill.
With every good comes some bad and AI is no exception. Today AI safety is as important a topic as AI itself. That obviously brings us to why? What harm can this self-learning algorithm do? Well, what if an AI is programmed to do something calamitous? Or worse, what if the AI is programmed to do something beneficial but it develops a destructive way to go about it? Thus, it is important to take into account that though the AI may be accomplishing its goals, our objectives may not be aligned with them.
This brings me to a contradiction. There are many who believe that super-intelligence by 2100 is inevitable and many who contradict by saying that super-intelligence by 2100 is impossible. But the fact is, it may take decades or centuries to fully develop, or might never even happen. AI experts are finding themselves in crosshairs about this and hence, no one really knows. People are concerned about AI turning evil or developing a consciousness of its own. But what we should really worry about is AI turning competent with goals misaligned to ours. Another myth is that robots may take over most jobs, causing most of this exponentially growing population to be rendered unemployed. But really, misdirected intelligence is what should really concern people as AI needs no body, it only needs an internet connection. Here, I would like to quote Elon Musk, an eminent researcher of the AI society, who said, “There’s no reason your Tesla won’t be able to fly in the future”. But, he also said, “AI is a fundamental risk to the existence of human civilisation”.
Thus, people with great insight and knowledge in this field see the benefits along with the fears of AI. AI is the future, but without AI security there will be no future.
Manipal Institute of Technology.
India