Burger King, one of the largest fast food chains in the world, has prepared a series of new prime-time television commercials in the United States, claiming to have produced them using artificial intelligence (AI) systems. Every commercial starts with the classic noise of modems of yesteryear to connect to the Internet, followed by a metallic voice that reads rather rambling and funny texts to describe burgers, such as: “A bed of lettuce you can sleep on, a bed of mayonnaise for an extra nap ”and“ Burger King chicken logo is the new potato ”. By Burger King's own admission, the commercials were actually conceived and created by real-life advertisers, in order to mock the high expectations and attention that AI has enjoyed in recent times, with sometimes questionable results.
As the Burger King marketing manager Marcelo Pascoa explained to the AdWeek site, the idea was to attract the attention of viewers by surprising them with something clearly unlikely, but still current and fun: “AI, bots, machine learning, algorithms for deep learning, blockchain and more – these are all topics we face in evaluating the future of marketing. However, we must avoid getting lost in the sea of technological innovations and fashionable words, forgetting what really matters. And this is the idea. Artificial intelligence cannot replace the great creative ideas that come to the mind of real people “.
Burger King's commercials, deceptive and apparently harmless, capture a central aspect in our relationship with new complicated technologies and still very blurred outlines, on which it is difficult to get an idea. In this sense, AI is an exemplary case: it is constantly talked about and it is what the largest investments in Silicon Valley and China are focusing on, yet there is still a big difference between what we read in the press releases that celebrate the great advances achieved by artificial intelligence systems and our daily experiences. There are AI who manage to beat the world champion of the very complicated Chinese game GO, others who manage the automatic driving systems of cars, but at the same time we have virtual assistants like Siri in our pockets that are sold to us as artificial intelligences, even if they are intelligent. they have very little and are almost totally useless.
The ambivalent perceptions that we have around AI, and on which the Burger King commercials play, derive precisely from a fundamental difficulty: to define artificial intelligence in a univocal and clear way. Behind these two words a large number of topics have been piled up ranging from pure computer science to neurology, passing through studies on how our brain works cognitively. Generalizing a lot, we can say that: artificial intelligence is the science that deals with how to create intelligent machines, and that has found in the possibilities offered by information technology the most practical and probable way to do it.
The concept of AI is deeply intertwined with the question of questions: How does human intelligence work? The discoveries on the mechanisms that govern the way we think, in fact, could lead us to develop the best AI possible; according to other researchers, the exact opposite could happen: we will develop a very powerful AI without understanding exactly how it works, and this will help us understand how brain and knowledge work. After all, already today there are many things that escape our understanding, in the scientific field, and that we are able to exploit perfectly to obtain concrete results.
Often for marketing reasons, and perhaps to attract the attention of some investor, researchers and companies promote systems as “artificial intelligences” that on closer inspection do not have much of “intelligent”. They do this by taking advantage of the very nuanced definition of AI, applying it as they see fit to their products. Away from their marketing executives and the ears of potential investors, virtually everyone working in the industry admits the same thing: at present, AI is stupid. But that doesn't mean it's useless.
There are very stupid people who excel at their jobs. Something similar happens with AI: we cannot say that they have a general and articulated intelligence like ours, but they have indisputable abilities. For example, they can analyze gigantic amounts of data that can form the basis for learning: through analysis, the machine itself gets an idea of the rules, whether it's a boxed game or how you cross the road, and then puts them practically. These and other solutions are part of research and development fields that have been greatly explored in recent years, and which concern deep learning, machine learning and neural networks. These are different things, but they can be used together to obtain more or less refined systems of artificial intelligence (here a minimal guide).
Advances in information technology, with increasingly powerful processors and computers, together with greater investments have led in recent years to a significant acceleration in the research of AI. The results obtained so far are very limited, but they bode well for the developments and innovations they may bring in the future. Ads, newspaper articles and marketing campaigns have skewed expectations: they have become much higher in a short time than is actually achievable by artificial intelligences. Also for this reason Siri, a robot dog or a car that drives itself appear much more stupid and banal than the technologies that make them work in reality.
In the essay “The farce of automation”, Astra Taylor questions the exaggerations surrounding the progress achieved with robots and artificial intelligences. In his analysis he introduces the concept of “fauxtomation”, that is the tendency of companies to promote the technological capabilities of their products by praising them in an exaggerated way. A microwave oven is defined as “intelligent” for the simple fact of being able to adjust the cooking minutes, after reading the bar code of the product it is about to cook and searching for it online. It's a way to sell more microwaves and it's pretty harmless, but it helps create a distorted idea of what we mean by 'smart' and where the technology is moving.
Automation, Taylor continues, is not as neutral and equidistant as one might imagine. In various industries it is used to justify unfavorable conditions for employees or to leave things as they are. The concept of AI is increasingly used as a threat to workers who demand improvements in their treatment, for example by giving them the possibility that soon a machine can do their job, at much lower costs. It is a much more dangerous aspect of “bogus”, which affects people's quality of life much more than a microwave sold as smart, even if it remains rather stupid.
The risk is that in this phase of transition towards real AI, which according to observers will be quite long (at least a decade), the out-of-reality idea of artificial intelligence promoted by those who develop it for commercial purposes could influence the decisions of politicians, and on the choices of public opinion themselves. Others might use “bogus” in reverse, for example to paint future scenarios where AI is necessarily a ruin, in order to reap short-term benefits in elections. A distorted idea about the actual capabilities of a technology can lead to bad and costly choices for everyone. A new technology is not inherently good or bad, but the use made of it can be.
