Embedded Files
Many professionals agree that Artificial Intelligence may come with either serious danger or a breach in ethical standards. While beliefs of Artificial Intelligence becoming self-aware and retaliating against humans may be rooted in fantasy, there are still pertinent dangers that AI presents to society. Common examples include Robocalls and computer viruses that aim to gain information on individuals, ultimately to scam them.
Image by Mohamed Hassan from Pixabay
"Tesla" by Automobile Italia CC BY 2.0
The fear of AI becoming too advanced and posing a threat to public safety increases with the continuing growth in technology. In fact, Greenwald (2018) quoted astrophysicist Stephen Hawking, who noted that AI, “could spell the end of the human race.” Elon Musk, who operates privatized companies that promote and explore the use of AI even admitted that researching AI is, “summoning the demon.” Further, Elon Musk’s application of AI to vehicles has already proven faulty with dozens of accounts about the auto-pilot system malfunctioning, in turn causing fatal accidents (Greenwald 2018).
Musk has also been the topic of ethically controversial AI attempts, one of which aimed to implant Artificial Intelligence into humans. This would essentially convert the human brain into a search engine, allowing humans to store mass amounts of information. While the possibility of this idea is questionable, many have still raised their concerns towards the ethical standards and restrictions that should be held toward Artificial Intelligence.
According to an organization called Geeks for Geeks, the five most pertinent dangers that AI may present are privacy invasions, the making of autonomous weapons, loss of human jobs, AI terrorism, and Artificial Intelligence bias, as depicted in figure 1.
Figure 1. Artificial Intelligence Dangers
5 Dangers of Artificial Intelligence in the Future. (2019, December 30). Retrieved from https://www.geeksforgeeks.org/5-dangers-of-artificial-intelligence-in-the-future/
To break this down, privacy invasions have the ability to occur daily, for example, when users unlock phones with fingerprints or facial recognition. Or for those who own artificially intelligent devices such as Siri, Alexa and Google Home, which have microphones constantly analyzing speech patterns and storing data, regardless of whether or not the device was activated. These systems contain an extensive amount of personal data, which, if fallen into the wrong hands, either through mismanagement or hackers, can cause serious harm.
"Facial Recognition Concept" by Nathan Gonzales
"SWORDS Robot" by US Military Licensed under CC0
Second is the idea of autonomous weapons, which is pretty self explanatory, however the idea basically holds that one could program weaponry to injure or kill others without a human controlling it. Essentially, this could even be programming self-driving cars to crash on purpose, or creating a robot that’s main purpose is to kill. While this isn’t something that has ever happened, it does coincide with the idea of AI terrorism, which has already become a danger to society.
AI terrorism is essentially when Artificial Intelligence is used alongside human efforts in order to complete an act of terror. This happened in 2016, where ISIS used a drone to drop bombs in Iraq ("5 Dangers of Artificial Intelligence in the Future", 2019). There are also fears that Artificial Intelligence mechanisms may be formed with bias coding. This could mean, for example, if a company uses an AI program in order to select candidates for a job, that AI program can be formed specifically to not choose certain classes of people, based on race, sex, class, etc.
This can also happen unintentionally, as the company Amazon released information stating that their algorithm for recruiting applicants had become, “biased against women,” due to the fact that it was based on the candidates hired over the last 10 years, which was majority male. In turn, the system thus began to favor males as it computed one of the qualities of being a good candidate as being male ("5 Dangers of Artificial Intelligence in the Future", 2019). Lastly is the potential for human jobs to become severely impacted by Artificial Intelligence, as human labor is far more expensive and generally has less capabilities than the work of AI.
Marine Corps photo by Sgt. Laiqa Hitt Licensed under CC0
A more niche concern presented by AI is the use of what are known as Deepfakes. Invented in 2014, Deepfakes are Artificial Intelligence programs that are used to recreate someone's face and put that face on someone else. Most deepfakes utilise GANs or generative adversarial networks which are able to not only classify data but generate or create images when provided with data or information. Some exceptionally strong programs can recreate the movement and likeness of a person from just one picture. The current main concern with these videos is that they are used for pornographic content by putting celebrities' faces in inappropriate contexts, which is misleading to online viewers and unfair to the privacy of famous people.
However, this technology could also be even more terrifying if used in malicious ways such as to recreate videos of political speeches or politicians, which could sway the minds of people unaware of the capabilities of Deepfakes. Fortunately, there are also many neural networks dedicated to identifying Deepfakes online now in order to determine if questionable videos are real.
References:
5 Dangers of Artificial Intelligence in the Future. (2019, December 30). Retrieved from https://www.geeksforgeeks.org/5-dangers-of-artificial-intelligence-in-the-future/References:
Greenwald, Ted. "Artificial Intelligence (A Special Report) --- what Exactly is Artificial Intelligence, Anyway? Everybody's Talking about AI these Days. here's what all the Fuss is about." Wall Street Journal, Eastern edition ed., Apr 30 2018, ProQuest. Web. 30 Mar. 2020 .
Libby, K. (2020, August 13). Deepfakes Are Amazing. They're Also Terrifying for Our Future. Retrieved October 28, 2020, from https://www.popularmechanics.com/technology/security/a28691128/deepfake-technology/
Guera, D. (2018). Deepfake Video Detection Using Recurrent Neural Networks. Retrieved October 28, 2020, from https://ieeexplore.ieee.org/abstract/document/8639163
Cover Image via wallpaperflare.com
AI could prove to be potentially dangerous in some cases. Some of these situations include when AI is used with the wrong intentions or when it is not programmed correctly. When it comes to using it for the wrong intentions, one example is how someone could program AI to steal others information like credit cards. Regarding improper programming a good example can be found in the article “What are the 3 types of AI? A guide to narrow, general, and super artificial intelligence”. The article explains, “Consider you ask an intelligent car to take you somewhere as fast as possible. The instruction “as fast as possible” fails to consider safety, road rules, etc. The intelligent car may successfully complete its task, but what havoc”(O’Carroll, 2017). This example in explanation describes a scenario where an AI system is programmed to complete a task, but could cause severe damage to objects and people in its path if not programmed with safety guidelines as well.
References:
“Importance of Artificial Intelligence in the Modern World.” Audiovisualaoce, 11 June 2019, from audiovisualaoce.com/2019/06/importance-of-artificial-intelligence-in-the-modern-world/.
O'Carroll, Brodie. “What Are the 3 Types of AI? A Guide to Narrow, General, and Super Artificial Intelligence.” Codebots, 24 Oct. 2017, from codebots.com/artificial-intelligence/the-3-types-of-ai-is-the-third-even-possible.
Page updated
Google Sites
Report abuse