The future of AI might be perfect. And that sucks!
If AI evolves to solve all our problems, what will we do with our perfect lives?
Google, Amazon and other gargantuan companies shouldn't aim for perfection, or making humans happy with AI… they should aim on getting the task done.
Artificial Intelligence, by definition is:
The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
In theory that's just amazing. Robots/Bots/AI will free a lot of our busy schedule, help us by doing menial tasks in our place, fast-track us to customer services without waiting time, solve problems and even have conversations with us (if you have no friends 😓).
But I'd like to introduce you to two possible futures:
- AI will understand what makes humans happy and ace everything so perfectly that we may look for ways to experience sadness.
- If AI will use data from humans, what could happen if the data is biased? Will AI be biased by human errors? (alert, this could be a bad bad future)
1. AI will understand what makes humans happy and ace everything so perfectly that we may look for ways to experience sadness.
I was recently in a meetup here in London with the title "Humanising AI". There I heard a very interesting talk by Matt Jones, who is the "Principal Designer at Google AI". For obvious reasons he can't talk much about his projects at Google, but he told us that he is currently working with Google Clips.
Here is a video of his talk, in 2015:
If you haven't heard of Google Clips, check this out. It's pretty much a small camera that uses artificial intelligence to take the right picture automatically when it sees something interesting. Annnnd at the same time the camera is learning. Google Clips is already aware that babies, cats, puppies make humans happy.
Happy/pleasing moments are what triggers this camera in order to capture what makes us happy.
Everything Google is making aims to get the human experiences right. You have no time to make a call? Google AI Assistant will assist and schedule stuff for you. Perfectly. Everyone saw this:
Even the "Mm-hmm" is perfect!
AI is learning human behaviour and the goal is to make us happy. And if we are a species with many things in common, will our experiences start to look all the same?
Hypothetically, websites performing perfectly because AI will navigate us exactly where we want to go; AI placing components in a website/app because it knows where we click the most; AI using colours that psychologically stimulate certain feelings; AI helping create products tailored to match exactly our behaviour and deliver flawless experiences. If this really happens, things will start to look the same…
Have a look at what you carry with you everywhere. Your mobile phone. How it used to look 15 years ago and how most phones look like now. Pretty much the same huh?!
If every AI nails a bit of what makes us happy… Then, what will make us sad?
If we get everything right, wouldn't we create a generation of anxious and frustrated humans?
💡Crazy idea for this crazy future: Will humans use drugs to experience sadness?
2. If Artificial Intelligence uses data from humans, what could happen if the data is biased? Will AI be biased by human errors? (alert, this could be a bad bad future)
Simple as this:
If an AI is trained to understand numbers, algorithms, and these are inputed by humans based on their data, therefore, can an AI make its calculations based on wrong human assumptions?
Take your time and think about this future…