The rising adoption of artificial intelligence has generated significant hype which, unsurprisingly, has led to misunderstandings about the technology itself and what it can and cannot do. We take a look at the six most common AI myths below.
Artificial intelligence has the potential to generate a lot of value for society and businesses operating within it. Self-driving cars and vehicles are being trialled, real-time translation systems are widely used, and smart technology is gracing our homes in a bid to cut utility costs. Throughout the day, most of us interact with some form of AI. Unfortunately, given its complex nature, misconceptions about artificial intelligence’s capabilities or lack thereof at this stage are rampant. Below we share six common myths.
#1: AI is 100% objective
When it comes to AI technology, data is king. Also required for its appropriate adoption and use are the rules and input from human experts. This is when the question of bias often surfaces. Human decision making in this area, like in others, can be flawed, shaped by societal and individual preconceptions that are often unconscious. The introduction of AI monopolies could become increasingly problematic in light of this. If you have a single AI system in charge, it presupposes you can create entire ecosystems based on prejudice. Both the availability of a diverse range of technological solutions as well as variety when it comes to the implementation stage and use of AI tools is paramount.
#2: AI will replace all jobs
The idea that AI will completely take over the job market is another myth with little to no foundation. While it’s clear that the technology has the power to transform the workplace and automate time consuming and mundane tasks, it’s unlikely AI will completely replace humans. The ability to control, analyse and manage large data sets can certainly augment more complex tasks. For example, AI can be used in healthcare for disease detection. However, just because AI will arguably change job profiles, doesn’t automatically imply human intelligence will become altogether redundant. In fact, our role in overseeing many of the systems in use, given the danger of biases discussed above, will become increasingly important.
#3: AI is already simulating advanced cognitive functions
The myth that AI is already an advanced cognitive system that can simulate human thought is largely down to Hollywood and popular film productions like The Matrix. Super-intelligent cognitive models that mimic human thought processes are no where near as advanced as one might think. The perils of certain AI developments have been communicated by many leading experts though, including former Google CEO Eric Schmidt. The business man shared his concern about Russia and China’s potential new positioning as superpowers. Imagining a world driven by the rise of cognitive computing is not difficult, however.
#4: AI and ML algorithms are easy to repurpose
Investment in AI and machine learning algorithms is increasing rapidly. Big leaps in their development have led to the assumption that once an algorithm is optimised for a single task, it can simply be repurposed. Contrary to common belief, reapplication is not as simple as one might assume. While machine learning algorithms have been reused in certain situations, this has not happened without strict human involvement. For example, while Google DeepMind’s AlphaGo program was able to beat a leading South Korean Go professional, its AI-powered camera had to be trained by professional photographers. That being said, multipurpose AI is likely to become more common in the future.
#5: Everyone is ‘using AI’
AI is among the most misappropriated terms in use today. A survey carried out earlier this year by MMC, a London based venture capital firm, found that 40% of start-ups classified as ‘European AI companies’ don’t actually use artificial intelligence and are simply cashing in on the hype. A total of 2,830 AI start-ups in 13 EU countries where surveyed. Larger providers can be just as guilty, many often outsourcing this area completely to another third party. Despite companies openly marketing their use of ‘AI’, the majority of firms actually work with machine learning software. Machine learning is one of the subfields of AI combining mathematical techniques to allow a system or machine to practice deriving information from the underlying database. ML techniques include both supervised and unsupervised learning.
#6: AI is not central to business strategy
Given the transformative nature of the technology in question, businesses ought to be able to identify its potential impact on their operations. This is why the implementation of AI, rather than an afterthought, should be a strategic initiative. Researching AI, understanding the applications for the business and then making a conscious decision to use the technology or not is essential.