Welcome to this week’s episode of the Self-taught Novice Podcast. This week’s episode is a debate on AI; as to whether the merits outweigh the demerits. It is an extension of last week’s episode on Algorithms and AI - its driven features and what it holds for Ghana and Africa. If you have not already, it’d be best to listen to last week’s episode of the Self-taught Novice Podcast so you can follow the discussion. Click here

So why say AI’s merits outweigh its demerits? From the last episode, we established that AI drives the industrial revolution of this age. With AI, industries have been made better and more effective and efficient and as industries run economies and human civilizations, this translates to the overall betterment of humanity. Take the example of the inception of the steam engine which initiated the use of machines in factories.

There was a lot of scepticism about the introduction of the steam engine with the fear that it would threaten livelihoods. The steam engine in the midst of all these went ahead to redefine manufacturing in its age, reducing human labour and increasing system effort. Similarly, in this age, AI is leading the industrial revolution. AI is connecting man with machine.

From this submission we are to accept that AI would definitely face scepticism and opposition in its initial stages. The danger with AI is that it is fast progressing at an exponential rate which we cannot keep up with. Google’s Deep Mind technology was able to beat the world’s best Go player with top speed after being fed tons of information. Go has multiple permutations, and there was no track of how the AI managed to win. However, this is being misconstrued as a benefit. All systems are susceptible to going rogue. AI is no exception; it is not perfect.

In areas where AI is used to manage essential services such as power generation and delivery, the control of fuel and water pipelines, and water delivery, with any mishap, the scale of tragedy would be catastrophic. It is questionable to abandon important aspects of our lives to a technology we cannot fully understand and keep up with. We must tread with caution.

 (There are two types of AI; Narrow AI and General AI. Narrow AI is generally not as smart and consists of the little things like how your phone camera auto focuses on objects when you’re taking a picture. General AI is more advanced and smarter. It includes IBM Watson and Google’s Deep Mind technology. It’s the kind of futuristic technology often portrayed in movies; the flying cars and the like. This conversation is focused on General AI.)

AI

is the replication of human intelligence on machines. We may not fully

understand AI, because we are still trying to understand human intelligence to

replicate artificial systems that can be run on human understanding. The

fundamental algorithms are still being understood. AI systems can be likened to

human beings, even though we may be aware of the use of our body parts, we

cannot fully predict our thought patterns. AI is not predictable and will solve

problems in ways we do not expect.

The

danger is that we cannot track AI. There is no traceable path as to how it

works and functions. Neural networks think on their own. AI is not controlled.

It seems the concept is that we abandon all our control to AI; that is the

expectation we have come to have. The threat of AI having access to all our

information is high. In spite of the numerous benefits of AI, if it goes

uncontrolled and unlegislated it can have devastating repercussions.

Legislation cannot keep up with AI. How can we keep up with something we do