Artificial Intelligence (AI) is often named in conjunction with Machine Learning (ML) as an increasingly prominent building block to a successful data strategy. Technology belonging to Google, Microsoft et al is progressively improving and becoming more accessible. AI is, in fact, a combination of Machine Learning techniques and Deep Learning that simulate human behaviour and thinking, so that machines learn to efficiently solve specific problems.
AI is no new phenomenon: The above time line depicts some prominent milestones: From entrepreneurs who adapted their product range according to the weather forecast, to Amazons first book recommendations and even self-driving cars.
According to Gartner (and the 3000 CIO’s that they interviewed) 37% of businesses have implemented some form of AI in to their business operations. This has grown by 270% over the past four years. Of course, the application and extent to which AI is used varies per organisation; 52% of Telecom companies work with chatbot technology and in the healthcare sector, 38% work with computer-assisted diagnosis assistance. Other areas in which AI is prominent are fraud prevention and recommendation engines in e-commerce. However, 54% of CIOs see the lack of knowledge in this area, within their organisations, as the greatest hindrance to growth. Therefore, creativity is required in order to train employees with a statistical background.
AI is already being used regularly in the following 10 fields.
AI is used in smart recommendation engines which make suggestions based on your behaviour, preferences, people with similar traits and product information. According to research, applying this intelligence to webshops leads to an increase of 5-15% turnover.
Virtual shopping assistants and chatbots considerably improve the user experience. Natural Language Processing technology is increasingly improving resulting in a more ¨human¨ interaction. A virtual shopping assistant is able to answer frequently asked questions, gives recommendations, provides detailed information regarding orders, applies discounts and sends reminders. On average 70% of shoppers exit their shopping basket before checking out, an automated reminder is usually very successful.
Fight against fraud
Credit card fraud (Dutch article) is a quite common occurrence within e-commerce. AI can help prevent it, based on patterns in data and behaviour. Another frequent issue is fake reviews: AI can help to identify these so that they can be acted upon.
Uber, in addition to logistics companies such as UPS make frequent use of AI for navigation, in order to calculate the smartest routes. They can optimise new routes quickly, so that changes can be handled with flexibility. Via the UPS MyChoice experience, customers are able to change routes and destinations, along with managing their delivery times.
Elon Musk recently presented the Tesla Bot, which will use the same sensors, cameras and AI as the Tesla cars. The robot is able to learn from new situations and will become smarter and more useful. Initially the tasks it performs will be quite simple, however the expectation is that the Tesla Bot will eventually be able to run an errand for you.
Presenting of the Tesla Bot, August 2021
Selecting and matching good candidates for vacancies has become an increasingly automated field. Much like a match on Tinder is becoming more and more AI driven. For context on this check out Black Mirror episode (S4, A4), in which everyone has an AI-deduced perfect soulmate. Also consider putting together a perfect dream team for a project within a large organisation.
In addition to the diagnosis apps, AI is used in the detection of cancer cells, chronic diseases and the efficient analysis of (lab) data. Philips (Dutch article) uses speech technology for e-Health solutions such as digital assistance and remote patient support.
Applications with relatively less complex calculations than ‘black-box models’ also play a major role in the healthcare industry. For example, Councyl (Dutch article) is developing a self-learning choice support system for healthcare experts in their field, with which no historical data is used. By presenting hypothetical choice scenarios to a group of experts, the so-called ‘weights’ can be determined for all consideration criteria. This way, healthcare experts know how many (%) of their colleagues make a certain decision with regard to a patient. The model is relatively easy to understand and thus effectively supports the choice of healthcare experts: Human advice in digitised packaging.
AI (Dutch article) helps to analyse the optimal time for sowing or harvesting by way of computer vision data and machine learning. AI bots also help with the harvesting, speeding up the process and making it more efficient than when carried out by people alone.
The interactive experience here is made as realistically ¨human¨ as possible using AI. The opponents (otherwise known as NPC’s, Non-Player Characters) behaviour is driven by responsive, adaptive and intelligent AI. The level of difficulty and gameplay can also be adapted to the individual player.
Artificial Intelligence is essential in the development of self-driving cars. All of the data being gathered by sensors, cameras, radars, GPS and cloud services is being converted into actions by AI. Infact, AI is present in all new cars today in the form of ABS and cruise control. Here Tesla presents their current AI projects.
AI determines the content that you personally receive on Instagram, Facebook, Snapchat and Tiktok. Facebook also uses a tool called ‘Deep Text‘ for this, so they can understand a variety of languages. AI also detects and removes propaganda and fake news, along with other undesirable content. Although this can regularly go wrong.
Just as AI is being used to develop deep fakes, there are other companies emerging that detect the artificially developed content. Take DuckDuckGoose (Dutch article) for example, a startup that exposes deep fake content by proving how a photo or video has been artificially developed, thus contributing to a safer digital environment.
AI helps with the targeting of a desired audience based on user data and behaviour patterns etc. Customer service requests are more frequently being carried out by chatbots in combination with people. AI can adapt real-time personalised expressions to the various users of websites, for example.
Creation and production
AI has created works of art, composed music and even developed computer programs. A year ago, this still felt very experimental, however its now getting better and being used more frequently. For example, product descriptions and other content is being effortlessly created by machines, much like highlight videos from sporting events. AI offers a valuable means of combining targeting with producing real time personal advertising content.
Unfortunately, AI is also increasingly being utilised in weapon systems for use in conflict. For example the ‘kamikaze drones’ that can stay in the air for a long period of time, waiting for a target to be recognised. There are also defense systems that react autonomously with each other, so things can get out of hand without the need for human intervention. Not exactly desirable.
Finally, a brilliant and especially fun example of an app using AI technology:
My personal favourite application is Brickit. An app that helps users with their lego building. Basically you throw your Lego on the ground, scan it with this app and you will receive instruction manuals for everything you could build with your Lego. Brickit uses machine learning to recognise all the Lego pieces at lightning speed. Cool!!
Discover all the possibilities of AI with a Data Sprint by Onbrdng. Book a free workshop now.
We support The Plastic Soup Foundation
Sumatrakade 91019 BJ AmsterdamThe Netherlands
4-5 Bonhill St, ShoreditchEC2A 4BX LondonUnited Kingdom