Could Super-Intelligent Artificial Intelligence Be Controlled? Scientists Say No
Artificial intelligence programmes may be seen in the headlines every day now, be it for a creepy AI generated image, or the perils of developing potentially sentient intelligence.
Sci-fi fanatics might think of advanced AI as an evil entity that controls everything. Of course, these beings that are powered by AI may be extremely strong and powerful. But will super intelligent AI be controlled by humans, if the need arises?
Controlling super intelligent AI
Unfortunately, researchers believe that controlling an extremely intelligent AI might prove to be difficult. The reason is quite straightforward - if AI surpasses human comprehension skills, it renders our processing capabilities inferior. If we are unable to comprehend the intelligence of the super intelligent AI, we may never be able to control it.
But all AI is programmed to be friendly to humans, right? Well, yes. But authors of a new paper suggest that we cannot program empathy towards humans in AI if we don't truly understand the scenarios that AI might come up with.
Also read:?A Terrifying Woman Named 'Loab' Keeps Appearing In AI Generated Images
"This is because a superintelligence is multi-faceted, and therefore potentially capable of mobilizing a diversity of resources in order to achieve objectives that are potentially incomprehensible to humans, let alone controllable," researchers wrote.
The team's reasoning is based on the "halting problem" that was proposed by Alan Turing in 1936. It attempts to understand whether a computer programme will reach a conclusion (and halt), or forever keep searching for the answer in a loop. Turing proved that it's possible to have this knowledge for certain programmes, but it's impossible to know the answer for every programme that could be potentially written.
Also read:?China's New AI Image Generator Censors Words Like 'Revolution,' 'Tiananmen Square'
In the same way, AI programmed to never harm humans may reach a conclusion (and halt) or not. Either way, it's impossible for humans to calculate and contain. Researchers propose that to contain AI, it could be cut off from parts of the internet or certain networks in order to limit its capabilities, especially if it's super-intelligent.
What do you think about the ethics of AI and its potential danger to humanity? Let us know in the comments below.?For?more in the world of?technology?and?science, keep reading?Indiatimes.com.
References
Nield, D. (2022b, September 16). Researchers Say It'll Be Impossible to Control a Super-Intelligent AI?: ScienceAlert. Retrieved September 19, 2022, from https://www.sciencealert.com/researchers-say-itll-be-impossible-to-control-a-super-intelligent-ai
