Machine Learning and Our Future

Machine Learning is all the rage these days. Be it computer vision, speech recognition, pattern matching or high-speed decisional capabilities, this century is the century of software. Like all technological revolutions, there’s potential for miracles and catastrophes.

Large corporations have started to realize^ that Machine Learning is a way to prevent smaller competitors^ from threatening them. This is because small companies can’t (yet) afford the huge infrastructure and Big Data investments that ML requires. It’s not surprising then that Microsoft, Google, FaceBook and others have open-sourced ML platforms, trying to attract developers and smaller companies to their ecosystems.

This post will touch on but a few of the changes we can expect in the coming decades thanks to the upcoming advances in Machine Learning. Looking at our history, we can see how the industrial revolution has supercharged our progress as a species. I believe that the Machine Learning revolution will make the industrial revolution seem like a snail in slow motion. This is both hopeful and scary.

The purpose of any post in the Futurology^ category is to launch a wild, boundless speculation regarding what he future holds regarding a certain concept. To get things going, here are some of the things I imagine can be accomplished in the near future (coming decades) by Machine Learning. Feel free to submit your own ideas in the comments below. With your approval I may integrate these in the article, giving proper credit.

  • Speech recognition is already quite advanced. In the coming decade, most day-to-day electronic devices will understand what humans speak. Additionally, these devices will form an interconnected sensorial mesh via the Internet of Things. Privacy will obviously be a major concern.
  • Without any prior technical knowledge, people will soon be able to talk to the robots that are about to enter our daily lives, both indoors^ and outdoors^. Countless jobs will be transformed or outright eliminated. Companies will jump at the opportunity of cutting costs. While some of this will have beneficial effects on some (company stakeholders for example), society might be negatively impacted as there will be plenty of those that cannot find a new job in a world that is increasingly robotized. Hopefully at least a part of the next generation of superrich few will empathize with the disadvantaged many.
  • Advanced algorithms are already able to take better (and much faster) decisions than humans in some fields (for example management of traffic, energy and bandwidth). This capability will expand to more and more areas. This development should not be confused with True Artificial Intelligence^, but will still mean that yet more jobs will be given to automated systems. Here’s for example how Google used machine learning to save a massive 15%^ in data center costs.
  • Companies that own data infrastructure will become dangerously powerful. Just look at how FaceBook allowed^ Russia to interfere^ in one of the most influential electoral contests in the world. Given the narrow difference between the candidates, it is quite possible that Russia’s influence (of which only a small part will ever be uncovered) will have been a decisive factor.
  • Governmental oversight could prevent a lack of balance in society, but strong lobbying from powerful corporations will continue to corrupt the purpose of government (in most countries, our representatives have long ceased representing us, if they even ever did).
  • The influence of ML throughout the economy means that society will have to find ways to protect those that are at risk of being crushed under the weight of the coming changes. In a way reminiscent of the industrial revolution, entire job sectors will become obsolete overnight, except that this time around the changes will come even faster and affect more people. Fortunately, we are also wiser and richer than a century ago so we are well positioned to find constructive solutions.
  • We may enter a period where creatives (artists) will again be in high demand. At least until True A.I. is upon us, machines still can’t create art. No matter how advanced an algorithm may be, the art it creates will still be a soulless mixture of unoriginal and random.
  • We run the risk of falling under total surveillance, aka Super Big Brother. This is much worse than what George Orwell could have even imagined when he wrote his 1984 novel. Super Big Brother doesn’t need humans to listen-in to conversations. It simply records everything (this is already being done, as the Snowden leaks proved). Then, somebody (like an oppressive authority that seeks to exterminate dissent) asks it to find certain information in text, audio or video recordings. If we now think that we have little privacy left, Super Big Brother will make things exponentially worse.
  • However, all is not so bleak. If good people act, there is another invention that will shackle Super Big Brother. That invention is open surveillance. All systems used for surveillance shall be based on open source software. All people being surveilled will be able to access their information and know why it was recorded (you were in a public space, you were suspected of a crime, etc.).
  • Government will put privacy back in the hands of the people. Because government has to be the people. And because lies and secrets never truly saved the world (although perhaps they helped in postponing major conflicts without, however, fixing the underlying problems). It will be openness and communication that will be proven to be the only way forward if we are to survive.

The Futurology Disclaimer: I do not claim that my ideas are original. I’m sure these suggestions are just scratching the surface of what can be achieved, but hopefully they’ve scratched enough to get somebody inspired to come up with more. I’m also sure many of these ideas are already being worked on by several organizations. If any of the ideas listed by anybody on this page are original and will benefit any organization, I expect credit to be given where it’s due.

Version history:

2017-11-23 – 1.0 – Written.

Share Anywhere:


1 Comments

Leave Comment

Your email address will not be published. Required fields are marked *