How AI and ML is helping to protect us from the dark side of the in...
October 9th, 2020
Algorithms you say. Are they relevant to your everyday life? Well of course not, right?
This is where you are wrong! They appear everywhere, and you may not even realise just how much they have an impact on your daily life.
“A process or set of rules to be followed in calculations or other problem-solving operations, especially by a computer.” – Oxford Dictionary
Have a quick search into a Google and you will find many different definitions yet no clarity in what they do. In the digital world, we live in today, algorithms are essential for even the smaller tasks in life such as going to an ATM machine for cash and scrolling through your socials.
The digital presence we live in has changed and as machine learning advances, we are relying more and more on algorithms. As more and more data is being created every day, the use of algorithms is rising. Some expert says understanding algorithms and the impact they have on human life goes far beyond basic digital literacy.
Facebook tailors’ updates and content to each specific user, which helps increase the user’s experience. How their algorithm works is by controlling the order your posts are shown to each user allowing them to see more relevant content.
Their original algorithm showed posts to be in chronological order this has now be altered to show published content that are based on your activity on the platform.
Instagram uses an algorithm which determines what posts you can see on your feed based on your activity. They have done this so you spend less time looking for posts that you might be interested In and enjoy your experience.
The way technology is going they may look into this again in the coming years and no doubt change it again. So, watch out as social media is forever changing and giving us new ways in which we view our feeds.
Nutella’s marketing campaign relied heavily on an algorithm to continuously produce new labels for its jars. These labels were created specifically from a set of colours the algorithms believe the consumer will find attractive. This limited-edition campaign was fed from newspaper articles and a campaign which then led to a consumer demand. This algorithm was unique Nutella’s manufacturer Ferrero worked with advertising agency Ogilvy & Mather Italia on the project Nutella Unica.
GCSE grades had to be delivered through an algorithm this year due to COVID-19. This algorithm used data from previous years as well as pupils predicted grades, due to pupils not being able to finish off their school year and sit their exams. Whether this outcome being fair or not left some students shocked.
The prime minister states the grades that were generated were almost derailed by a mutant algorithm.
Used heavily in today’s world, where would we be without Google. Have you ever thought of how it comes up with results you are searching for?
Algorithms are set with a search engine to sift through thousands if not millions of webpages to generate content for the user. Having an account will allow a more personalise result due to the knowledge they have on you and your history.
What are the ethical concerns?
It’s the first thought when someone thinks of technology/robots, how are they advancing rapidly ‘They will take over’ and as that is possible it is important to stress that we create them and in order for them to work they will always rely on some kind of human maintenance.
With an algorithm, it does not operate in a single field or produces effects in a single manner which raises the question of who is imputing the data into them?
Inadequate data training can be a cause of algorithmic bias. If the data used to train the algorithm are more favourable on one particular type of person the predictions from the model will then be favourable to another. Bad input within the making of an algorithm will produce bad output.
An example of bias in AI is Amazon. Back in 2018, they abandoned a project to build an AI recruitment tool, which was found to discriminate against female candidates. The engineers built the program from resumes from a 10-year period, which were predominantly male. As there were no data on the gender split in its engineering department, the female candidates got filtered out. When someone applied who did not fit the profile the algorithm had learnt from historical data it filtered them out.
It is not the algorithm that is biased the bias comes from the data that is being inputted. Algorithms that are built, how they are developed are based on human decisions or on the data that reflects a decision.
Bruno Maisonnier an AI expert explains that there are 3 major sources of bias in AI algorithms
It is safe to say that algorithms can shape your next purchase, what you listen to or even your next step. However, we choose to live in a world where we will all at some point be involved in some algorithm and for that reason, it is necessary for us all to get to know and understand it.
As Machine Learning advances it is imperative for us all to get to know and understand it. Why? The use of algorithms is only likely to grow, who said you had to be a software developer to have an understating of algorithms?
So, it is important to know that even the little tasks you do daily could potentially be based on algorithms.
Without us even knowing algorithms are here to stay and will further enhance.