By Ariosto Colón-Zolikoff ESMol S6ENA
Recently, during a televised interview with BBC Panorama, Brad Smith the president of Microsoft warns that if laws are not put in place to protect the public from Artificial Intelligence (AI) then humans could be living in a world like George Orwell’s surveillance society in his novel “1984”. Smith also predicts that without legal regulation around the world this type of controlled totalitarianism could be a reality on Earth in the near future – even as soon as 2024. AI combines huge sums of data with algorithms which allow it to ‘learn’ from the patterns of data. As AI begins to play a greater and more prominent role in society, the world will be forced to question whether current AI, as well as the AI of the future, will be a help or hindrance to the survival and prosperity of the human race.
The European Commission (EC) have recognised the threat that AI poses and on 21 April 2021 announced that regulations to prohibit AI which can lead to “physical or psychological” harm are under serious consideration. The use of AI biometric identification systems for policing and law enforcement will also be included in this legislation. New legislation would address the “potential high risks [AI] poses to safety and fundamental rights”. The European Union’s foundational principals are human dignity, freedom, democracy, equality and respect for human rights. The EC, which is at the forefront of this movement, propose to enact laws to ensure that these principles are upheld in relation to AI.
Artificial Intelligence is already demonstrating its potential to harm the integrity and well-being of people in all domains with respect to racism, invasion of privacy, severe job losses and misinformation. There is also the question of the looming moment when Artificial Intelligence surpasses the cognitive capacity of humans entirely.
Currently, AI has demonstrated on numerous occasions that it is easily possible for bias (whether unintentional or not) to manifest itself as discrimination or persecutory tendencies. When utilising Artificial Intelligence for firing, hiring and for making arrests, users may make unjust decisions based on potentially incomprehensive data which the AI programme has ‘learnt’ and from which it makes conclusions.
An example of data bias is a landmark study conducted by the US National Institute of Standards and Technology (NIST) in December of 2019 which found “facial recognition software” (a form of AI) to accurately recognise white males the most often and black females the least often. This startling revelation prompted an investigation which discovered that the data set for the technology included fewer black women and many more white men. Using AI to assist with policing which unconsciously recognises one type of person more accurately than another can increase the number of unfounded and incorrect arrests as a result of data which is unrepresentative of a whole society.
Despite these distinct drawbacks, there are those who argue that AI is essential for maintaining control of societal order, ensuring quality of life and reducing costs while improving efficacy of the world wide labour force. However, a combination of revelations of bias and inadequately advanced technology among other reasons have contributed to the fact that the use of Artificial Intelligence is not yet so widespread.
The current lack of extensive control and regulation of Artificial Intelligence within institutions and companies could have monumental challenges and dangers for mankind. The public have ready access to current AI technology such as ‘deep fakes’ (a software used to create fake realistic videos and audios) which can be utilised maliciously for a range of purposes from spreading misinformation (e.g. in elections) to destroying reputations.
Although many believe the threat of AI in computers to be negligible, when Artificial Intelligence is implemented in other important applications (such as pacemakers, aeroplanes, bank accounts and cars) the ramifications of a security breach or even a simple mistake could be catastrophic. In addition to this, the use of extensive AI surveillance has raised serious questions of whether this is a breach of privacy and human rights. This criticism has had a large focus on China who monitor their citizens extremely closely.
Modern-day AI (which is in the stage commonly known as ‘weak AI’) is still only able to perform relatively narrow activities (e.g. to play chess or to recognise faces). The more worrying problems are set to arise with the development of so called ‘strong AI’ which will outperform humans at most tasks.
However, unlike in science fiction, AI is unlikely to takeover the world but has the potential to have misguided objectives with parameters that deviate from what humans consider ethical or correct. On this topic, the entrepreneur Elon Musk stated: ‘AI doesn’t have to be evil to destroy humanity. If AI has a goal and humanity just happens to come in the way, it will destroy humanity’. This is why all AI must be closely regulated to ensure the safe operation of such powerful technology.
With the rise in the uptake of AI, many workers will be displaced from their jobs by an AI counterpart. The professions deemed most threatened by an increase in the usage of AI are: economists, accountants, teachers, surgeons, police officers and journalists. This takeover of jobs by Artificial Intelligence will leave millions unemployed while giant corporate software developers gain large sums of money because of the ubiquitous use of their product.
Some may argue that Artificial Intelligence will increase the efficacy and productivity of workers by eliminating jobs which require analysing and working with large pieces of data. This solution will come at the cost of leaving millions upon millions unemployed and having to either spend extra time retraining or to simply live off the support money of governments.
Artificial Intelligence may play a key role in our future but unless stringently controlled, AI could become a menace and a threat. Furthermore, the possibility of AI created with unintentional or intentional bias could have an extremely negative and dangerous impacts on the population of the world. What’s more, an introduction of AI in the workplace will have a detrimental effect on the work environment resulting in the loss of many jobs. The
In conclusion, despite various counter arguments, such as how AI will make working more efficient, scenarios where negatives, such as millions becoming unemployed, basic human rights being violated, and potential lives being lost, clearly outweigh the near-sighted naive benefits. The invention of new AI may very well be the undoing of the human race and the prevention of such an event is paramount to our survival. The European Union seem to be leading the world in the right legislative direction.