OpenAI and Microsoft's Plan to Create a Futuristic AI
OpenAI is an AI research company based in San Francisco and was founded most notably by Elon Musk and current CEO Sam Altman. Microsoft is investing $1 billion into the corporation and will provide OpenAI with cloud computing services. Both companies aim to develop new technologies together, specifically a groundbreaking concept – Artificial General Intelligence (AGI) – highly autonomous systems that outperform humans at most economically valuable work.
However, the research company has been a part of the scene since 2015. For instance, they developed an AI which can beat professionals at Dota 2 as well as text-generation systems which can write anything from short stories to news articles. Likewise, AI is being employed and developed in the world by many other different companies as well. Other systems are known to compete and ruin masters at chess, spot cancer, trade stocks, and play Breakout and Super Mario like champions.
Even so, as you might notice, these are all confined tasks. Although artificial intelligence is meant to think on its own after being trained in environments, they are often successful only in those particular environments.
Can AGI defeat this problem? Of course. AGI aims to be flexible enough to surpass all of these AI across an extremely broad array of tasks. Exactly like a human. As such, a system capable of this can change the way we live and usher in a new stage of human development, but it still poses a serious threat if not handled properly.
A successful AGI can solve human problems like diseases, hunger and corrupt political systems. Centuries of technological advances can be brought to humanity immediately due to the extreme speed of autonomous research. Imagine the revolutionary catalogue of inventions arriving all at once. The benefits would be inconceivable.
Nevertheless, the security threats which come with artificial intelligence as capable as AGI are terrifying. Although it may not necessarily lead to the AI suddenly developing hostile intentions, or creating a new race of further AGIs which eliminates all of humanity, consequently recreating a real-life dystopia like the Terminator. Instead, the problem arises when instructions we give to the AGI are unclear and vague which could lead to it following orders in ways we didn’t intend. Or worse, a minority could develop an AGI with malicious intentions in mind.
Avoiding these issues may be one of the biggest challenges OpenAI could face in their pursuit of a successful AGI.
OpenAI’s main goal is that their research should benefit all of humanity. Yet, until this year, they were a non-profit company and intended to provide research on AI safely and democratically. Their shift to a for-profit organisation could be seen as abandoning these morals.
Despite their transition, they do want to uphold their initial aims by operating as a capped-profit entity. Simply, this means that the company can pay shareholders up to 100 times in return for their investment, but any profits above that margin go to the public.
It is still unclear how fast such technology which still seems science fiction can be developed. Some experts say it’s right around the corner, possibly only a decade away, whereas others predict that we’re still pretty far off. Nevertheless, if a safe version of an AGI is produced humanity will be launched into the most exciting era of growth and progress. Let’s just hope we’ll be able to experience it as well.