The Risks of AI : A concern for the future

lmost every website on the internet today possesses AI-generated content in one way or another. From art such as AI music to AI-generated scripts of television shows, AI is now dominating most of what we see in the digital world. Given the recent improvements by OpenAI in its AI chatbot, ChatGPT, it has now opened the eyes of the world to the risks and dangers AI brings with it.

In the past, whenever I would hear about artificial intelligence and futuristic technologies, the first thing that came to my mind was software driving machines to take over the majority of human work. From cleaning cars to packing items in factories and even driving cars themselves. But while all of this might still be under development, AI has already become an essential part of our lives. Social media algorithms and voice assistants are simply AI technologies that we have been using unknowingly. AI was thought of as something that would make human life a lot easier, but with its benefits come risks that can completely change the world and how we perceive our future.

With the launch of generative AI, we can only observe the beginning of what AI can do. It has been observed that school students now rely on ChatGPT to do most of their homework. The work that used to take students hours to complete is now done in a matter of seconds, and it doesn't require much effort except for writing the prompts. This is concerning because these students will be the future leaders, and if they rely too heavily on AI to do their work, it may hinder their ability to develop essential skills and knowledge.  Furthermore, there is no guarantee that the information generated by AI is correct. AI often provides answers with confidence, making it hard to distinguish between truth and fiction. 

AI experts who recently wrote an open letter about how AI could lead to the destruction of humanity have shown concerns about people relying on AI for health advice, emotional support, and decision-making in real life. These experts predict something scary about the future of AI if the development of new AI technologies continues. It could lead to the destruction of humanity. The risks of present-day AI systems include information biasness, joblessness, and more. However, with the development of advanced technology, companies, governments, and independent researchers can develop powerful AI systems that can control everything from business to warfare, even things we don't want them to. Experts also suggest that these systems can resist our attempts to stop them. While this cannot happen with present-day AI systems, it is a possibility with those currently under development, such as the AI system AutoGPT. These systems can produce actions from prompts, generate computer programs, retrieve information, use apps, create new ones, and even self-improve. However, they are still caught in an endless loop for now. There is still a lot of effort needed to develop this technology that would improve itself, as stated by Connor Leahy, the founder of Conjecture, an organization that aims to align AI with human values. He also believes that currently, these systems don't work, but someday they surely will. Criminals could provide prompts like making money, and these systems could end up looting banking systems.

Another factor that proves how dangerous AI can be in the future is the misbehaviour of AI technology. AI technology systems are developed on networks where they learn new skills by analysing data from the internet, and they can learn more than even their creators, which might cause them to exhibit unexpected behaviour. Recently, an AI bot even passed a captcha test by pretending to be a visually impaired human.

While all of this might happen in the distant future, there is still a need for those responsible for regulating AI to realize the harms that present AI systems pose and ensure effective regulation.