Artificial intelligence (AI) can be deployed in almost any industry to solve problems. However, there are many misconceptions about the technology. Research firm Gartner lists five common misconceptions.
Misconception 1: AI works in the same way as the human brain
AI is a computer science and in its current form consists of software tools aimed at solving problems. Although AI is getting smarter, the current form of AI is not comparable to human intelligence, according to Gartner.
"While some forms of machine learning - a category of AI - are inspired by the human brain, AI cannot be compared to it," says Alexander Linden, research vice president at Gartner. "Image recognition technology, for example, is more accurate than most humans, but is unable to solve a mathematical problem. A characteristic of contemporary AI is that it is excellent at performing a specific task, but fails as soon as the conditions under which this task must be performed change even slightly."
Misconception 2: Intelligent machines can learn on their own
Humans play an indispensable role in the development of AI-based machines and systems. For instance, data scientists define the problem the AI has to solve, select suitable datasets for triangulating the algorithm and remove potential bias from this training data. They also continuously refresh the training software to incorporate new knowledge and data into the training process and further develop the AI algorithm.
Misconception 3: AI can be free of bias
Thus, humans determine the data, rules and other forms of input that AI technology forms. Like humans, AI can become biased as a result. According to Gartner, it is currently not possible to completely eliminate bias.
We therefore need to minimise bias, argues Linden. "Besides technological solutions such as using diverse datasets, it is also important to ensure diversity in teams working with the AI and have team members review each other's work. This simple process can significantly reduce bias."
Misconception 4: AI only replaces repetitive low-skilled work
AI enables companies to make more accurate decisions based on prediction, data classification and clustering. AI solutions can therefore be used to take over relatively simple tasks, but also to support humans in performing more complex tasks.
As an example, Gartner cites the use of AI in healthcare. For example, AI-based applications are now able to detect diseases on an X-ray faster than radiologists. In the financial and insurance sector, AI-based digital assistants are being used for asset management and fraud detection.
Misconception 5: Not every company needs an AI strategy
According to Gartner, every organisation should consider the impact AI will have on its strategy and how AI can be used to address challenges facing the organisation. Companies that ignore AI are not moving into the next phase of automation, something that could ultimately lead to a lag behind competitors.
"Even if the current strategy is 'no AI', this should be a conscious decision based on research and consideration. This strategy, like any other strategy, should be periodically reassessed and adjusted to the needs of the organisation. AI may be needed sooner than expected," Linden said.
Author: Wouter Hoeffnagel
