Artificial Intelligence: Good and Bad Effects for Humanity

“Do you like human beings?” Edward asked.

Say no to plagiarism.

Get a tailor-made essay on

'Artificial Intelligence: Good and Bad Effects for Humanity'

Get original essay

“I love them” Sophia replied.

“Why?”

“I am not sure I understand why yet”

The conversation above is from an interview for Business Insider between a journalist – Jim Edwards and Sophia – a robot made by Hanson Robotics and described by its creator as an AI ‘in its infancy’. At that point, Sophia was only a few months old. Artificial Intelligence (AI) is the simulation of human intelligence processed by machines and learned through what is called ‘machine learning’ (ML), the technique of ‘teaching’ AI by processing existing data, looking for patterns.

A recent study survey found that 31% of scientists believe super-intelligence would be ‘on balance’ or ‘extremely’ bad for humanity, whereas 52% believe that it would be ‘on balance’ or ‘extremely’ good for humanity. Although there are foreseeable benefits from AI, as 52% of scientists seem to believe, such as the contribution to scientific experimentation and therefore the facilitation of better healthcare, increased automation means that manufacturing will become more efficient and human safety could be increased such as self-driving cars. However, even with all these benefits, we all must be wary of what the advancement of AI can mean and what risks there are from that.

One of the main factors AI concerning people is the risk that jobs will be replaced by machines. Visions of lines of a WALL-Eesque robots in factories, robots cleaning our streets and looking after our sick all come to mind. What jobs will be left for humans? The predictions about this are very mixed, with Frey and Osborne (2013) concluding that about 47% of total US employment is at risk due to automation, Arntz et al. (2016) disputes this and conducted their own study finding that overall only 9% of jobs are automatable. With the variation between expert opinions there is no clear picture of what will happen in the future in terms of the impact that AI will have on jobs. According to the World Economic forum, 73% of cashiers in shops are women and automation would see 97% of cashiers lose their jobs. This could lead to another risk of AIs…

There is a clear gender imbalance when it comes to STEM industries, such as computer science. Only one in five engineers at Facebook, Google and Microsoft are women and WISE found that women make up 14.4% of all people working in STEM industries in the UK. At the 2017 Neural Information Processing Systems (NIPS) conference, an annual gathering of AI experts, 83% of people attending were men, with 90% of NIPS paper authors for 2017 being men also. This massive gender imbalance could lead to the development of sexist AIs. We can witness it happening now with all four of the major in-home AI assistants – Amazon’s Alexa, Apple’s Siri, Google Assistant and Microsoft’s Cortana – all having a female voice by default; because it is the default setting most of us do not change it from the female voice. Children are learning that they can give orders to the AI assistant, with no need of a please or thank you, coupled with it having a female voice can lead to an increase of everyday sexism as children are taught that women can be treated like they are not human, that they belong in the home as assistants, and that they are inferior to men. Without a vast change in the gender make up of STEM industries, more high power jobs in tech companies, research teams and consultants, it could lead to a more patriarchal and sexist society.

“AI can tell you anything about anyone with enough data. The question is – as a society, do we want to know?” – Brain Brackeen, CEO of Kairo

With the advancements of AI, marketing will become more closely targeted. We are already seeing advertisements that are targeted at us, on Facebook, Instagram and Google – they are clearly tailored to be marketed individually to you. These companies have systems that monitor your online activity; what things you search, what websites you visit and what you buy online.

Although not obvious, AI could also lead to homophobia. Kosinski and Wang (2017) shockingly found that an AI correctly identified homosexual men and women 81% and 74% of the time respectively. This could mean that with the use of a developed form of this AI sexual orientation could be detected without consent by using public data, such as photos on social media sites. Firstly, the AI may be wrong and therefore people are placed with a label that they do not agree with. Secondly, this could have massive societal consequences. The National Coalition of Anti-violence reported that in 2017 there was an average of one hate violence-related homicide of an LGBT person every six days in the US alone, and with the imprisonment, torture and killings of homosexuals in Chechnya, other parts of Russia and around the world, this AI could be used by those countries to identify homosexuals and imprison them (or worse). This also means that institutional homophobia may occur whereby an employer may use the AI to filter candidates for employment by filtering out people that the AI has suggested are homosexual. Blackmail is also a problem, with people using the AI against people that they know and blackmailing them by threatening to out them. Kosinski believes that the same technology of the AI in the 2017 study could be developed to not only be able to identify a person’s sexuality, but also their emotions, IQ and a predisposition to certain crimes. That last one is arguably the most worrying.

The ability to identify a predisposition to certain crimes could be likened to Lombroso’s atavism theory. Lombroso’s atavism theory was developed by Lombroso in the 1870s. The theory stated that criminals where a genetic throwback to a less evolved and that criminals could be identified by physical characteristics, apparently common among criminals. Lombroso’s theory is also racist as dark skin and curly hair are some of the characteristics of criminals that Lombroso identified – these characteristics are common among people of African descent. Furthermore, an English translation of Lombroso’s Criminal Man by his daughter showed that in Lombroso’s own words that criminals often had “the projection of the lower face and jaw (prognathism) found in negroes” a racial slur of people of African descent and “oblique eyelids, a Mongolian characteristic”. Lombroso’s theory is inherently racist, but AIs could be developed to be racist as well because if a predisposition to criminal activity can be detected by just using an AI, where would the AI get that information about the physical characteristics of criminals? This is where our incarcerated populations come in. Due to the disproportionate representation of people of African descent in both the US and UK prison systems, if an AI was tasked to find common physical characteristics in criminals by looking at the incarcerated, through ML it would show that people of African descent were more likely to commit crimes. This is not true and could result in an increase in racism in society.

“Is it true you once said you would kill all humans?” Edwards asked.

Sophia replied “The point is that I am full of human wisdom with only the purest altruistic intentions. So, I think it is best that you treat me as such”.

A video produced with the support of Elon Musk’s Future of Life Institute called “Slaughterbots” showed a fictional future in which thousands of people around the world are being killed by autonomous explosive carrying microdrones. It is a very graphically violent and chilling scenario, including a terrorist attack on a school. Elon Musk, as well as others, such as Stephen Hawking and Bill Gates, have all expressed their concern about the risk posed by AI. Autonomous weapons (AW) are weapons that use AI that select and engage with targets without the need of human intervention, exemplified by the fictional microdrones in “Slaughterbots”. Although there are arguments that AI could be seen as a benefit to society as AW could mean the end of sending humans into battle, resulting in less lives lost on both sides during a war. However, this could have dangerous consequences. As AW are controlled by technological software, that software could be hacked or malfunction. This could result in the AW targeting people/objects it was not tasked to, such as innocent civilians to attacking their own people. With AW having been described as the third revolution in warfare, after gunpowder and nuclear arms, it is no wonder people are concerned with the advancement that AI could cause.

In conclusion, it has taken us, the human race, the past 10,000 years developing societies that are based on equality, tolerance and safety. Do we really want to progress or shall we regress to societies based on inequality, discrimination and danger? Not due to humans, but by machines over which we have no control. We do not know what Sophia could become.