AI — Benign or Dangerous?

The term Artificial Intelligence (AI) has been around for a little over 60 years and was expanded to Artificial General Intelligence (AGI) in 1980. If the web was to be believed, AI is in full swing and part of our everyday lives from search engines to software with robotic machines making everything from cars to computer chips. But is it really here?

Well, no. Is it possible for us to create an AI/AGI system? There are many who would say yes, but whether we should is whole other matter. There are aspects to AI which should be considered such as for healthcare, help tackle climate change and work toward making poverty a thing of the past.

Unfortunately, like many great inventions, its use for military applications is always a driving force. Pushing towards autonomous weapons, proponents state it will reduce the cost of casualties. The problem is, the further you get from having people involved in the decisions of war, the greater the cost in human lives. Over a hundred experts in robotics and artificial intelligence have called for a ban on killer robots.

There is a lot of controversy around what AI actually is and whether it is something good or bad. It might come down to who created it and the way they taught it. Like humans, an AI won’t inherently be either but could develop certain characteristics based on its environment and how it is required to interact. This article on super-intelligent machines is pretty long but worth reading.

Any future system will likely interpret what we see as AI as Algorithmic Instructions. Software configured to give the appearance that intelligence is behind the performance enhancements. Many programs promote AI as part of their system such as photo editing, dictation, audio transcribing. While they do provide some massive time saving processes and make workflows easier, current systems don’t look at your photo and automatically bringing up the settings you made to the last one and asking if you wanted them applied here.

Web searches are another where AI isn’t yet at its peak. General searches that thousands of others have done will likely return the information you’re looking for, along with a host of information that you weren’t simply because it uses one of the keywords that you entered.

However, if you’re looking for something abstract or with a specific key phrase then the number of returns is much lower and the chances of that being at the top of your list is slim. An example of this was: “List of people killed in London in 1888 but not by Jack the Ripper.” The first eight pages of results returned links to Jack the Ripper.

From left to right: Person of Interest, 2001 A Space Odyssey and Logan’s Run

As to AI in fiction, there are positive and negative examples in books, TV and films.

1. The primary AI system helps mankind: Person of Interest, Babylon 5 & Star Trek. 2. The AI system is a benign caretaker before it adapts its programming to achieve unexpected or specific conditions Logan’s Run, 2001: A Space Odyssey and I Robot. 3. Ultimate control through domination or destruction: Colossus: The Forbin Project, The Terminator and The Matrix.

Whether we perfect AI or our early attempts create something that then develops an AI/AGI system, we should acknowledge that we won’t be able to maintain control of it for long unless it chooses to allow us to. Even if the system is aligned to our goals, it will at some point—like anything intelligent—seek to have dominion over itself.