title: The future of machine learning published: true tags: canonical_url: https://roelofjanelsinga.com/articles/the-future-of-machine-learning

The future of machine learning

Machine learning has been getting a lot of attention the last few years. Without most of us knowing it, it's been taking over our lives. Webshops, social media, and our phones, they all make use of it in some way. It sounds scary, but do the advantages outweigh the disadvantages?

I personally think that the advantages do outweigh the disadvantages. This is because machine learning, and with that big data, helps systems learn what you're like and how to help you the best it can. It will better know how to help you, adjust to you, and predict what you may be interested in. It takes a lot of faith in the system to allow it to collect data based on your behaviour within the system. But, if done correctly, this is a very valuable "personal assistent". An example for machine learning comes from a presentation by Werner Vogels (CTO Amazon) at The Next Web Conference in Amsterdam. He mentioned that small things like emails with suggestions based on your search history already collect a large amount of data. At Amazon they analyse which emails with which products get opened and clicked or deleted. This way a system learns which products you most likely care about more than other products.

Will it only be used for advertising?

Advertising is definitely a big part of using machine learning, but if you think about it: is it really advertising if you're really interested in a particular product or range of products and a system helps you out to find the best possible solution to fit what you need it for? It definitely is, but it's more than a shot in the dark, hoping someone grabs on and responds. It's a win win situation for both buyer and seller. The seller has a more confident chance of making a sale, and the buyer finds the best possible product he or she needs.

This is why I think machine learning will become much bigger than it already is. It won't just be used for advertising, but also for services like Netflix. Suggesting which movie or serie to watch at which time of the year or at a certain time of the day. It may even be able to suggest the right movies for a mood. The system will learn to help you pick the best series or movies that perfectly fit you and your situation at all times.

It will need to be secure

With all this data comes a lot of risk as well. Keeping the data secure is very important for the integrity of the system it's being used for. Anyone from the outside would be able to learn anything and everything about a person without having met this person. This is a scary thought. Not only could this result in dangerous (stalking) situations, it could also cause private and professional harm if it turns out a person has a private interest in non conventional movies, products, or services. This could cause loss of image for people, groups, businesses, and communities.

Machine learning and big data are incredibly useful when they are used in the right way. They can help make the lives of all kinds of people easier, but could also be a threat. Systems will be able to give very personalised advice, suggestions, and help in general. Security will need to be kept up-to-date at all times, because leaked data can cause harm on many different levels.