🚨OpenAI to Counter ‘HUMAN EXTINCTION RISK’ (this is NOT clickbait) OpenAI, the creator of ChatGPT, plans to counter what they call "the vast power of super-intelligence [that could] lead to the disempowerment of humanity or even human extinction." And they admit that "we don't have a solution for steering or controlling a potentially super-intelligent AI, and preventing it from going rogue." Remember, OpenAI are the makers of arguably the most advanced Generalised Artificial Intelligence alongside Google's DeepMind. If this was a Hollywood script, it could pass as a Terminator spin-off or a new prequel to Battlestar Galactica! Remember when @elonmusk warned that “the pace of progress in artificial intelligence is incredibly fast —it is growing at a pace close to exponential. The risk of something seriously dangerous happening is in the five-year timeframe. 10 years at most.” WHY BUILD ‘SKYNET’ DESPITE EXTINCTION RISKS? This is the first thing that comes to my mind. Why do we want to make super-intelligence? Well… I'm not alone. In April this year, more than 1000 people in the field of AI including Elon Musk, Emad Mostaque (Stability AI), many researchers at DeepMind and various professors said “the big players are becoming increasingly secretive about what they are doing, which makes it hard for society to defend against whatever harms may materialize." So we’re at a crossroads, either we: 1. Pause innovation in regards to super-intelligence 2. Keep innovating as fast as possible 3. Somewhere in the middle, slowing development to avoid things getting out of control Option 3 obviously seems like the most logical solution, except that we’re human and humans are stupid. We’re too short-term oriented, focusing on financial returns at the detriment of long term risks. Another issue is capitalism, meaning that any slowing of innovation in the US will simply give competitors like China the edge. WHY IS IT HARD TO CONTROL? DeepMind and OpenAI have been dealing with a problem they call alignment of AI, which can give unexpected results. In a nutshell, these systems are 'taught’, not programmed (think of a program as a script). So the AI is given feedback that is positive if ir does what we want, but if it does something wrong, it is given negative feedback. It’s essentially the same as training a dog or a small child. Unfortunately, the issue with alignment is that sometimes this doesn't work and the machine does something unpredictable. Now mix this with super-intelligence that's basically smarter than us, and the mistakes could be the launch of nukes or the release of an infectious disease. SO WHAT WILL OpenAI DO? Well they have a plan and it goes something like: We will train AI to correct the alignment of AI and make it as smart as possible. Yep, you read that correctly: AI will fix AI’s problem. Connor Leahy, an AI safety advocate, said the plan was flawed because "you have to solve alignment before you build human-level intelligence, otherwise by default you won't control it. I personally do not think this is a particularly good or safe plan." IS ALL A.I. DANGEROUS? The answer is obviously no, AI will change the world we live in for the better in many ways, more than we can imagine. The question is though, will the benefits outweigh the risk? Here’s a list of current risks of A.I based on today’s limited applications: 1) Autonomous weapons 2) Social manipulation 3) Invasion of Privacy 4) Misinformation and fake news 5) Bias and Discrimination 6) Security and Internet 7) Job Displacement… And the list goes on. MY THOUGHTS: AI worries me more than Russia, more than a Nuclear war, more than pandemics and more than global warming. At the same time, it’s one of the most exciting technologies I could even imagine. I think the ship of regulating AI before it gets out of control has sailed. Human nature will force us to innovate as fast as we can despite the risks to our very existence. People will not pay attention to this either, not until it's too late. I hope I'm wrong... DOES A.I. WORRY YOU?