Will Artificial Intelligence Make Us Lose Our Jobs?
By Vitalii Duk, Data Scientist at dubizzle
Humanity experienced several technological revolutions, which mainly led to an increase in productivity, efficiency and overall improvements in our lives. Usually, when certain jobs were eliminated, the new ones, which never existed before, were born out of newer innovations and technologies. As an example, digital marketing specialists didn't exist before computers were invented.
According to Gartner, AI will create 2.3 million jobs by the year 2020, while eliminating 1.8 million . So, we should think about AI as a new tool that will help us to explore new frontiers and embrace even more opportunities.
In the meantime, it is good to familiarize ourselves with the current capabilities of AI, it’s pros and cons, and possible applications in different businesses, so it will be easier to adapt to the new AI technologies.
Is AI capable of taking over the world?
Despite fears that Artificial Intelligence could be taking over the world and control over human beings, in fact, the current state of technology and algorithms is far from being even close to human intelligence. While scientists are working towards creating a general-purpose AI which potentially can be capable of solving different kinds of tasks, most of the current successful breakthroughs are just complex mathematical models trained to perform only one specific task well. For example, Alphabet's subsidiary DeepMind created a neural network that plays video games, is entirely self-taught and can beat any human being. However, the software is not yet advanced enough to tackle some of mankind’s most challenging problems.
Typically, when most people refer to AI, it is usually meaning a particular Machine Learning algorithm which was created at the intersection of computer science, mathematics, and statistics, allowing computers to learn complex patterns based on provided data.
Most of the popular algorithms and techniques in Machine Learning were developed from around the1960’s to the 1980's. For example, a very first prototype of Deep Neural Network, which is currently a state-of-the-art approach in many Machine Learning tasks, was first successfully used back in 1965 . Where Artificial Neural Networks are supposedly created as an approximate model of a human’s brain, they are still based on very naive assumptions of how our brain works, creating a lot of limitations.
Why has AI become so popular?
There are two main factors which led to the popularization of AI and Machine Learning.
First, the technology enabled huge amounts of data to be stored at a lower price. This development was crucial as data is the primary key to success that makes all the algorithms work. Most of the algorithms simply cannot extrapolate and provide meaningful results with a small set of examples, they usually require millions.
Second, computational resources became cheaper and available for a general audience, often referred to as the ‘democratization of computational resources’. The power of existing algorithms was unlocked by modern Central Processing Units (CPUs) and Graphics Processing Units (GPUs), which became more efficient and less pricey.
These two factors embraced a lot of new opportunities that were not available before. Companies started storing a lot of more data and all the information that previously was very expensive to store was now processed to generate valuable business insights. On the other hand, it became possible to use some of the existing Machine Learning algorithms, which depended either on computational resources or a number of available data points.
How does dubizzle use Machine Learning?
AI and Machine Learning enabled technology companies like dubizzle to flourish, harnessing the power of big data sets and insight-driven solutions for end-users. dubizzle always aims to provide better seamless experience to its users.
Among many projects where dubizzle uses Machine Learning, here are the primary areas:
- Personalizations and recommendations: Machine Learning algorithms are capable of analyzing terabytes of user data, which allows us to provide everyone with personalized search results and feeds, taking into account users’ interests and preferences.
- Spam and fraud detection: While working behind the scenes, dubizzle’s AI-powered spam detection system is preventing irrelevant ads from being posted on the platform and allows us to process thousands of ads per second.
- Image recognition: Powered by state-of-the-art Deep Neural Networks, image recognition allows dubizzle to simplify the User Experience by detecting placing immediately after user upload an image.
- Forecasting: At dubizzle, Machine Learning is widely used among Strategy Department in order to forecast many of business KPI’s and improve decision making.
Plans for the future
For an online company like dubizzle, using Machine Learning technologies is crucial to succeed in the competitive technology market. Adoption of AI systems within the different businesses will continue to grow rapidly, as it can help companies to develop exceptional and outstanding features, which will distinguish it from other competitors on the market. Staying ahead of the technology curve when it comes to AI means companies will deliver increasingly customized and tailored features to end-users, ultimately creative a competitive advantage in the technology industry.