Artificial Intelligence (AI) has occupied the minds of forward-thinking people since last century. Lots of sci-fi novels or movies have featured it in one form or another form, but now we have come to a stage where it is no longer a fantasy.
Business leaders feel like their companies will sink in the ocean of innovations if they don’t apply Machine Learning, Computer Vision or Natural Language Processing in their products or operations. However, it is necessary to keep in mind that an investment in AI only makes sense from a long-term perspective, and it is not something that will just automatically have a “wow-effect” on customers. Moreover, implementing it may cost more than the return on investment due to the high computing force and complicated infrastructure that is needed.
However, when it is applied wisely, the benefits can be huge. For example, for Google Plus, the application of machine learning improved important metrics, like expands, reshares, plus-ones and comments per read, comments and reshares per user, etc. Efficiency was increased by 2%, which may seem like nothing, but when applied to a huge organization like Google, it results in hundreds of millions of dollars in gains.
There are a lot of successful examples of its application in different industries, and we have collected several use cases to inspire you to think about how your business can be improved by Artificial Intelligence.
Recently, we published a review of healthcare software trends and examples, and Machine Learning was listed as an efficient tool for the detection of rare diseases.
In a recent study conducted by the University of Bonn and the Charité - Universitätsmedizin Berlin, 679 patients with 105 types of rare diseases were diagnosed using a neural network that combines portrait photos with genetic and patient data. The scientists “trained” the software using around 30 thousand pictures of diseased people, and the software was capable of filtering out the genetic factors that could be causing the illness. By doing so, the time for data analysis was decreased, and the software discovered causes in cases where it was previously impossible.
The application of Artificial Intelligence software and Machine Learning solutions in this domain has started a lot of discussions. There are three main components necessary for AI/ML application in this field: lots of data, effective process pipelines and design models and security domain experts. And many companies have been criticized for lacking at least one of those. Many companies have been criticized for lacking at least one of these requirements.
However, a Texas Tech University research team has proved that Deep Reinforcement Learning, another AI-based technology, is highly efficient for the detection of phishing websites. This cybercrime, which aims to obtain the user’s data like banking details or credentials and passwords, is usually conducted via emails, calls or messages. To get to this step, “phishers” create malicious websites, where “victims” are asked to enter the above-mentioned information. Reinforcement Learning can use repetition to teach computers how to recognize these websites and prevent phishing attacks.
Banking and cybersecurity are logically connected, as cybersecurity is an indispensable part of online banking and fintech. However, quick identification of fraud is not the only benefit. Embedded AI tools help to customize user experiences according to individual customer’s behavior and habits. For example, by obtaining data from different sources bank AI can sort expenditures depending on the counterpart of a transaction and improve financial planning services, making the whole customer journey more smooth.
Another way to engage AI in online banking services is by involving it in credit scoring procedures. When a customer makes a credit or loan request, AI/ML checks not only the basic data but also data from extra sources, like utility payments or mobile phone use, in order to make faster real-time decisions.
One of our partners, FIDO solutions, is one of the most vivid examples of successful implementation of this technology.
Though we have already claimed that AI is not an extremely spectacular technology, entertainment is one of the fields where its performance may be more obvious. Alongside the more conventional personalization of user experience, like, for example, the predictions of user preferences implemented by Netflix, AI can also be used for content creation.
20th Century Fox has collaborated with IBM to create the trailer for its new horror movie Morgan. They have trained the Watson AI software on parts of 100 other movies of the same genre that were categorized based on their audiovisual and compositional features.
As soon as the system was trained, they gave it 10 recommended scenes and the full-length Morgan movie, and in 24 hours (compared to weeks required for manual trailer production) AI compiled the footage that you can see below.
Last but not least is on the opposite side of the tech business world. Widespread AI development has also become important in the world of art.
Art curator and gallery director Aidan Meller has introduced an AI robot called Ai-Da that creates abstract images from shapes and colors on the Cartesian coordinate system. By giving Ai-Da a more or less human appearance, Mr. Meller engages the public in a dialogue concerning the future and outcomes of AI technology.
In all of the above cases, we can see that AI, with its rather stealth intrusion, is changing daily lives without claiming its presence. Thus, it is important to stay in touch with this technological development and see how it can be viably and sensibly implemented into your product in order to not miss the right moment and keep up with the competition.