Artificial intelligence – Solving human problems
90 % of all available data were generated only during the past two years
From autopilot to pilot
One of the most common images of AI is the self-driving car, a phenomenon that has grown exponentially in recent years. But for pilots, assistive technology has been around for a long time. The autopilot, which helped aircraft to fly level and straight, was invented as early as 1912.
The big difference is that autopilot technology eases the workload for pilots, whereas the self-driving car eliminates the human factor from the equation. In self-driving cars, AI makes decisions based on learned patterns previously observed in traffic, together with the current situation around the vehicle. The technology that enables it to do this is called “machine learning”.
Different kinds of machine learning
When companies mention AI, they are often referring to machine learning (ML) technology. ML can be described as a self-learning programme that changes dynamically on its own over time by means of data analysis and seeks an optimal outcome, such as winning a game of chess or driving a car without crashing.
Learning can take place in two ways – supervised or unsupervised. Supervised learning is used when there is a clear picture of the objective, for example diagnosing changes on the human skin. Unsupervised learning is used when we want to identify contexts that are not clear to humans. Today’s enormous supply of data is driving the development of AI, and as models are improved, the number of possible applications increases and the potential for new companies improves.
OK Google, what’s traffic like right now?
Even today, Google has a good sense of our road traffic situation thanks to machine learning. With millions of users sending data about their speed, position and destination, Google can create a picture of the current traffic situation, including which roads should be avoided and how best to reroute the user.
Other companies that have implemented AI technology are Amazon (to shorten the time for product deliveries), Facebook (for face recognition), John Deere (to estimate future harvests) and Spotify (to create personalised customer playlists). This huge potential is attracting companies as well as governments to invest in AI and related research.
A growing multi-billion dollar industry
Revenue from AI-related software is expected to increase from USD 9.5 billion in 2018 to USD 118 billion in 2025, an annual growth rate of nearly 40 per cent. Most AI patents are held by the world’s tech giants, with IBM and Microsoft standing out as number one and two in the number of patents. One of China’s explicit goals is to become the global innovation centre for AI by 2030. Europe seems to have fallen behind, but the European Union has established ambitious AI investment goals and has appointed expert groups in order to catch up with other continents.
The AI landscape today
Today’s AI landscape can be divided up between developers and users: companies building the ecosystem required for AI applications and companies that use AI technology to update their business models and create competitive advantages.
Corporate acquisitions within both categories have increased significantly. Such acquisitions were previously made mostly by major technology companies, but have now spread to other sectors that are trying to catch up by purchasing technology developed by smaller companies. In the past decade, nearly 650 AI-related corporate acquisitions have been carried out, with around 140 of them just this year (for more details, see The race for AI).
Industrial revolution 4.0
As more processes are automated and handed over entirely to AI, the need for AI technology will increase. In many sectors and industries, the AI revolution is already knocking on their door today. According to PricewaterhouseCoopers (PwC), about 30 per cent of today’s jobs are expected to disappear by 2030 as a result of AI. But in the short term, the need for highly educated people with degrees in quantitative subjects will lead companies to increase their number of employees in order to implement various types of AI technology. Companies that wish to retain their competitiveness may be forced to invest in this technology to avoid being overtaken by competitors that take advantage of its economies of scale. As a result, AI technology may impact the bottom line of more and more companies. But considering the gains – in both human and economic terms – it seems like an investment in the future that is worth living with.