Introduction
In the past 10 years artificial intelligence has become more prominent in the technology industry, with sophisticated voice recognition being implemented in now common house hold products, like the amazon echo, or be it deep machine learning being used in a variety of tasks, as complex as self-driving cars. Fundamentally artificial intelligence is a program “that makes it possible for machines to learn from experience, adjust to new inputs and perform human-like tasks”. (SAS 2018) as a result of the self-adjustment and human like nature of the programs some people fear of a robot uprising fuelled by a superior artificial intelligence, as depicted by movies such as the terminator and ex machina. However, are there fears justified in the 21st centaury or is artificial intelligence actually a threat.
Before the issue is discussed fully the type of threat that is going to be discussed must be defined. The definition of a threat as stated by the oxford English dictionary is as follows “A
person or thing likely to cause damage or danger. “As consequence of this definition it is plausible to question whether artificial intelligence is a threat due to its inclusion under the “thing” stated in the definition, yet in what way is it a threat, for this report threat will be discussed in regards to being a threat to a human’s life, as depicted by many Sc-fi movies.
Artificial intelligence in society
The role of artificial intelligence in society is the largest that it has ever been, it is used in a wide variety of applications and use cases from industry, self-driving cars and data analytics to virtual assistance and more novel uses. The figure below from an article shows the way in which machine learning has many use cases. (O’Reilly State of Machine Intelligence 2.0, Shivon zilis December 10, 2015)
As shown above it is clear that machine learning has its most uses within many different branches of industry, not only just in the home. It is estimated that 33.2 million smart speakers were sold in 2017 (Canalys, 2018 march), as this number demonstrates artificial intelligence is becoming more of an accepted technology within society. In a survey carried out by ARM relating to artificial intelligence the most well know was the apple assistant Siri with 77% of the sample group being aware of its AI features. It is the wide spread nature of the technology and the uncertainty as to what Its full potential is that constitutes to the fears that people have over it.
Current concerns about artificial intelligence
In its current state the largest concerns of artificial intelligence are the risk of it being hacked, weather this be to listen in on your conversations or hacked and your private data stolen. There is some evidence that suggest that your smart home devices, such as those that implement machine learning can be hacked in order to listen in on your conversations. This is supported by the claims made by ex-CIA member Edward Snowden, who in an interview with the BBC stated “Nosey Smurf is the 'hot mic' tool. For example, if it's in your pocket, can turn the microphone on and listen to everything that's going on around you”. However, this isn’t concerns against the Artificial intelligence its self, more so the organisations and governments behind them. The scale of this concern is further reflected by a survey that was carried out by the CPU company ARM. In this survey a group of 3938 people were asked questions concerning AI, one of the questions asked about concerns relating to AI and security. When asked as to whether they were worried about AI machines being hacked and personal information being stolen 85% of the survey groups said that yes this is a concern they have. (ARM,2017) Moreover similar claims of these concerns have come from notable academics such as Steven hawking, who in a news article stated the following” The development of full artificial intelligence could spell the end of the human race."
How could it be a threat.
In the world of film and media, the dangers of artificial intelligence are displayed to be those of robots gaining conscious to such a level that they have the ability to cause harm and kill. The name for such risks is coined by the term “Existential risk from artificial General intelligence” this is the hypothesis that artificial intelligence could one day become so powerful that it results in the destruction of the human race or a global catastrophe. However, in its current state the chance of this occurring is unrealistic. Even though these risks are unlikely, there are ways in which current artificial intelligence could be a risk to life.
One of the most topical ways in which it could be a risk is through the development of driverless cars by large technology companies such as google, Apple and Uber. The development of driverless cars has been ongoing now since 1925 when radio producer Houdina radio control created a radio-controlled car called the “American wonder” that they drove down Broadway, however this test eventually ended in disaster when it crashed into another vehicle. (Discover magazine, 2017). Since then the development of driverless cars has come a long way, with then becoming so complex that they can learn for themselves as stated in a report by Neil McBride,” Using learning algorithms, they will build knowledge bases of road conditions and learn to managing unusual and exceptional conditions “. It is this fact that the car learns for its self that scares people and makes them be perceived as a threat to life. On march 18th 2018 a 49-year-old woman was killed in a fatal incident involving a self-driving car developed by Uber, this is the first and only accident of its kind involving fully autonomous cars. The accident occurred due to a conflict with the way in which it was developed, at the time of the accident the car was in “prediction mode”, Sebastian Thrun, a Stanford professor who used to lead Google’s autonomous-vehicle effort stated that “this is the hardest mode to develop, our perception module could not distinguish a plastic bag from a flying child.”(The Economist 2018) It is this uncertainty and trust that is put into these systems that results in them becoming a risk to life. Even though there has recently been a death as a result of driverless cars, there is data that suggest that people are willing to trust driverless / autonomous cars, in the ARM survey those who took part were asked the following question “Next year / In the next decade, would you trust an autonomous car to drive your family if their accident rates were demonstrably better than human drivers? “the response to this question was remarkably high with 55% saying that they would trust them now and 77% saying that they would trust them in 10 years’ time.
Even though fully driverless cars are the projects of some of the worlds largest companies there are aspects of them that are prevalent in main stream cars already, particularly features such as autopilot found in cars from the company Tesla.
Reference List:
• Shivon zilis, (2015), State of Machine Intelligence 2.0, O’Reilly media: https://www.oreilly.com/ideas/the-current-state-of-machine-intelligence-2-0 [Online]
• ARM, (2017). AI Today, AI Tomorrow: http://pages.arm.com/rs/312-SAX- 488/images/arm-ai-survey-report.pdf[ online]
• T.S, (2018) Why Ubers self-driving car killed a pedestrian, The Economist: Can be found online at https://www.economist.com/the-economist-explains/2018/05/29/why-ubers-self-driving-car-killed-a-pedestrian
• Carl Engelking,(2017),The “driverless” car ear began more than 90 years ago, discover magazine December edition
Source reliability
Type of source Source Author Reliability Score of reliability (1-10)
Web Article Artificial intelligence, what is it and why does it matter. SAS Analytics company.
Dictionary definition Oxford English dictionary University of oxford The Oxford English dictionary is an internationally renowned and cited dictionary written by one of the world’s top universities 10
Interview Edward Snowden BBC interview Peter Taylor The reliability of this source is highly questionable due to the source of the claims, hence why it is only being used as a quoted example. 4
Survey ARM; AI today, AI tomorrow. ARM/NorthStar A joint venture between a technology firm and a global insight company. However, the sample size of 3938 isn’t large enough to be representative of the global opinion on AI 7
Magazine Carl Engelking