The effect of technological illusions on people’s perception of reality

Technological Illusions - The effect of technology on people's perception of reality

Article content

Computerisation, which began in the 1990s, has propelled humanity into an era where working and interacting with technology on a daily basis is common and natural. Artificial Intelligence answers our questions, and the Internet is seen as an endless source of information. While one may think that the development of technology helps us to understand the world around us, there are phenomena that show how often our intuitions fail. Technologies, which at first glance are simple and obvious, can hide paradoxes and illusions, which may be more difficult to spot, as well as to understand, than it seems. This article will explore three interesting phenomena: the ELIZA effect, the Moravec’s paradox and the Streisand effect. Each of these shows how technology can change our perception of reality, affecting how we see machines, data and information. Exploring these phenomena will provide a different perspective on the development of technology and help us to use it more consciously. 

ELIZA effect 

In the 1960s, Joseph Weizenbaum at the Massachusetts Institute of Technology developed the ELIZA programme [1]. This programme was one of the first chatbots – it naturally mimicked a normal conversation. Despite the simplicity of the algorithm, which created responses based on the keywords entered by acting according to predetermined patterns, many users of the programme reported that they got the impression that Eliza really understood them. Thanks to the clever selection of answers, users were able to be highly engaged in the conversation, satisfied that the interviewer understood them and was paying attention. The creator himself was surprised at how convinced people were that Eliza was a human being, not a machine.  

It is from this chatbot that the ELIZA effect, the phenomenon of the tendency for humans to attribute to machines, programmes (including AI) the capabilities of understanding, empathy and intelligence, i.e. to anthropomorphise them [2], got its name. Examples of this phenomenon include the appearance of “hello” or “thank you” messages on ATM and self-checkout displays, which are pre-defined texts to be displayed rather than an expression of gratitude by the machine; or communication with voice assistants – thanking them, saying “she” about the Alexa assistant, which despite using a female voice still remains a genderless algorithm. The reason behind this effect can be attributed to our nature – everything that is human seems familiar, closer, less frightening, which can be seen, for example, in the way ancient deities are depicted and compared to humans and animals, attributing weather phenomena or elements to them [3].  

Such bonding with, sometimes very complex, technologies allows one to overcome fear of novelty, encourages interaction and builds attachment to the product being used. At the same time, this effect can cause an overestimation of the capabilities of a given algorithm (due to the assumption that the machine knows and understands more than it actually does), excessive trust in the information received, or an inappropriate treatment of the creation as a human being, e.g. by treating a chatbot like a therapist or marrying AI [4]. 

Moravec’s paradox 

Another interesting phenomenon takes its name from the Canadian scientist Hans Moravec, author of works on technology, futurology and transhumanism. In 1988, he, together with Rodney Brooks and Marvin Minsky, formulated the statement: “It is comparatively easy to make computers exhibit adult level performance on intelligence tests or playing checkers, and difficult or impossible to give them the skills of a one-year-old when it comes to perception and mobility” [5]. It implies that tasks that are considered difficult, requiring knowledge, intelligence and logical thinking, are relatively easy to solve using AI, while those activities that we consider simple and natural – walking, recognising faces and objects or motor coordination – are very challenging and difficult to implement in machines.  

Researchers attribute the reason for this paradox to the human evolutionary process. Human motor skills developed over millions of years and were essential for survival and slowly but continuously improved by natural selection. The human brain has had plenty of time to assimilate and adapt to activities such as grabbing tools, recognising faces and emotions, walking or motor coordination, so they are automated at a deep level; we perform them without conscious effort. On the other hand, abstract thinking, mathematics, logic are relatively new abilities, not rooted so deeply and requiring conscious intellectual effort. Because these abilities are not ingrained so deeply in human beings, it is easier to apply reverse engineering to them and implement them in the form of a programme. In addition, computers are most effective at mapping logical, schematic processes with specific steps. For these reasons, we already have programs that are superior to humans when it comes to complex calculations, chess, simulations, but when it comes to mobility, coordination, object and face recognition, or other “basic” activities that we consider natural and simple for a child of just a few years old – the development is very slow. It is only recently that the amount of data and technology has allowed a gradual development in this area, as shown, for example, by the robotics design company Boston Dynamics [6].

Streisand effect 

Another phenomenon presented in this paper is the Streisand effect. According to this phenomenon, the more one tries to remove or censor a piece of information on the Internet, the more publicity and interest it receives. The effect owes its name to Barbra Streisand and the situation that occurred in 2003, when photographer Kenneth Adelman took photographs of the California coastline to document the progressive erosion [7]. These photographs were made public on a website dedicated to the subject of coastal erosion. Coincidentally, one of the photographs showed Barbra Streisand’s residence. She sued the photographer for invasion of privacy, demanding damages and the removal of the photograph as she did not want anyone to see it. However, it turned out to be quite the opposite – she lost the lawsuit and had to reimburse the photographer, and not only was the photograph not removed, but it received even more publicity and many more views than before the whole situation.  

This effect can be attributed to several factors, mainly based on human psychology, the role of social media and the general mechanisms of online information circulation. People are very reluctant to endure any restrictions imposed on their freedom, including access to information. Often, in situations of enforced censorship, people deliberately act out of spite – they want to get as much news about the “forbidden” information as possible, are willing to share it and spread it further. The “forbidden fruit” effect works in a similar way – by attempting to hide the information, it appears even more interesting and intriguing, even though without the attempt the message would probably have been disregarded. Nowadays, because of the ease of access to information and the multitude of different media, news is widespread and can quickly become viral, attracting huge audiences. The Internet has also changed the perception of various content. In theory, the fact that any user can save and share content makes it impossible to remove something from the Internet once it has been posted there. Given also how quickly the media seize on and publicise instances of censorship, it becomes quite obvious why an attempt to hide or cover up something usually ends up having the opposite effect. 

There are many examples of the occurrence of the Streisand effect. In 2013, after her Super Bowl performance, singer Beyonce’s publicist deemed one of the photos particularly unfavourable and attempted to remove it from the Internet. The effect was exactly the opposite; the photograph became considerably more popular than it had originally been and also began to serve as a template for internet memes. There are also many examples of the Streisand effect from the world of technology. In 2007, a user of the Digg website revealed that the Advanced Access Content System (AACS) copyright protection system used in HD DVD players could be cracked with a string known as 09 F9. Representatives of the industry using this protection demanded that the Digg post be removed and threatened legal consequences. As a result, a great deal of discussion took place on the Internet, and information about the code (which for a while was referred to as “the most famous number on the Internet”) spread heavily and was reproduced in the form of videos, t-shirt prints or even songs [8]. 

Summary

The phenomena discussed in the article show that although technologies such as Artificial Intelligence and the Internet are powerful tools, they have the potential to distort human perception and create misleading impressions. It is easy to fall into the various traps related to technology, which is why awareness of the phenomena mentioned is important, as it allows for a more critical approach towards interaction with technology and information, a better use of their potential and their healthy and sensible application.  

References

[1] https://web.stanford.edu/class/cs124/p36-weizenabaum.pdf 

[2] https://modelthinkers.com/mental-model/eliza-effect 

[3] https://builtin.com/artificial-intelligence/eliza-effect 

[4] https://www.humanprotocol.org/blog/what-is-the-eliza-effect-or-the-art-of-falling-in-love-with-an-ai  

[5] https://www.researchgate.net/publication/286355147_Moravec%27s_Paradox_Consideration_in_the_Context_of_Two_Brain_Hemisphere_Functions  

[6] https://www.scienceabc.com/innovation/what-is-moravecs-paradox-definition.html  

[7] https://www.forbes.com/2007/05/10/streisand-digg-web-tech-cx_ag_0511streisand.html  

[8] https://web.archive.org/web/20081211105021/http://www.nytimes.com/2007/05/03/technology/03code.html