Artificial Intelligence (AI) is an emerging technology, and as with every emerging technology, it is always accompanied with both excitement and worry. AI holds great potential in solving real world problems. However, this potential can cause damage if not properly managed by society.
AI is an umbrella term used for any computer software that has the ability to reach a goal through trial and error, this goal is set by a software developer. Since the goal is set by a human and the solution is provided by a machine. AI proves to be a very powerful tool that can be used by humans to create a positive or negative impact. This being said, I believe that modern use of AI is of negative impact to society.
Nowadays when the question “Is AI a threat?” is posed, society tends to think of the threat in a science fiction manner, i.e. Super intelligent cyborgs take over the world aka the singularity threat. Yet, this is far from current reality. Currently AI has the “intelligence of an earth worm” and is still decades away from becoming something human like. Society is still ignorant to the fact that AI is already impacting their everyday lives, especially when it comes to social media platforms.
Social media is powered by AI and the "AI in social media" market is projected to become more than a $2.1 billion market by 2023. This number just proves how synonymous AI is with social media. AI is used in multiple ways: From keeping the users engaged through addictive design to showing the users ads that are tailored just for them. Although the latter has some controversy around it, I find that the addictive design aspect is what really impacts our society.
Addictive design translates into AI algorithms that are implemented on social media to keep the users engaged. These Algorithms tend to show the user more of what they like and less of what they don’t. This creates confirmation bias. Confirmation bias is when a person gathers or interprets information in a way that conforms with his own pre-existing beliefs and ignores contrary information. Right now, all social media users are guilty of confirmation bias without their knowledge and that’s because of AI. The AI algorithms creates what is know as an “echo chamber” where the social media user is held in a loop of posts/thoughts that are adjacent to their thoughts, with no exposure to posts/thoughts that oppose them. This causes polarization and might lead to extremism of thought. this is very dangerous to society, especially in a political context. A study on polarized topics on social media quantitively showed the existence of echo chambers in a political debate in Brazil concerning the impeachment of Brazilian President Dilma Rousseff.
Due to the polarized medium in social media, misinformation is quickly shared within the polarized communities. This gave rise to the use of Social Bots in a malicious context. Social Bots are AI powered users that are designed to mimic human posting in large quantities. The designers of these Social Bots might design them to poison political debates and introduce faulty information for their own political gain. This allows the developer to effectively manipulate social media (hence, society) by spreading misinformation and low-credibility content through the use of AI.
Aside from the polarization of social media through the use of AI. Addictive design powered by AI technologies pose another threat to society which is “mass influence”. Tik Tok is a prime example of an app that uses AI and has influence on masses. Tik Tok is a platform that allows users to scroll through short videos, automatically showing the user more of what they like and less of what they dislike, making the app free of human influence over what they see. The ease-of-use Tik Tok introduced made the app so successful that “by December 2020 the app had been downloaded 6 billion times”. Of the millions of users Tik Tok has on its platform (with an average use of 52 minutes per day) no one has control over what they see. The control is all in the hands of the “Tik Tok algorithm”. Through filtering information and tweaking the algorithm to fit a certain agenda, Tik Tok could potentially have tremendous power of influence on societies all over the world. This has led the US government to lead an investigation into Tik Tok’s use of its algorithm.
Although Tik Toks influence over societies is something that is to be wary of, governments still have authority over apps, and they can regulate the ethical use of AI. The problem becomes a threat to societies when the governing body is the one responsible for such AI use (influencing and tracking users in the millions). This is becoming a reality in China where the Chinese government is using AI to govern their citizens. And although it might be effective, it raises a lot of concerns about the impact of AI on countries.
The Chinese government has put forth plans to become the world leader in AI by 2030. They have also voiced their opinion in regulating their people through the use of AI. In the starts of 2020, the Chinese government started experimenting with a digital system that assigns people colours using a risk scoring AI, and this score controls their access to facilities and public transport. This was implemented in hopes of containing the COVID-19 spread throughout their nation. This algorithm creates precedence for a country wide digital system that is controlled using AI. The next step that the government can go forth with is a digital system that is controlled using a sophisticated algorithm that assigns the Chinese people with a social score according to their “social standing” in the eyes of the government. Again, using AI as a tool, they can control people based on their social standing. This will have a severe impact on society’s day to day life and how governments influence their people.
Moreover, the government might use AI in a way that imposes a threat to society by using algorithms to track people all over the country. The Chinese government has been accused of just that after some leaked documents showed that the government is using “facial recognition technologies” to track people (China’s Muslim ethnic minority in Xinjiang province) and to always keep surveillance on them. This imposes complete disregard to the Chinese populations privacy and raises flags of racial apartheid that is enabled by the power of AI.
Yet another cause of worry is China’s “smart court system” in shanghai. A system where AI-generated sentences are given to defendants. This introduces biases and unfairness to the judiciary system. AI is trained on human biases, and they tend to amplify those biases, so having this AI system judge without showing the decision-making process to the defendant or showing what biases might have been implicit, makes it near impossible for an appeal and even more so for a fair trial outcome/sentencing.
Human biases tend to get amplified by AI as AI is trained on information provided by humans/developers. In 2015 Amazon discontinued a trained AI system that was used to help recruit employees. The decision to discontinue the AI came from an unexpected bias in its decision making. The AI taught itself to prefer men over women by scoring resumes lower if they included the word “women’s” along with other words. This was because the AI was trained on data provided by humans and the data showed a causality between being a female and not getting the job. Hence, the algorithm enhanced this bias. This goes to show how AI could impact society. If amazon had not realized this discrepancy, they could have used this recruitment system as an “objective” recruitment system and the bias would have never been disclosed to the public. And this Amazon example proves how the Chinese “smart court system” introduces a significant negative impact on society and decreases the peoples trust in a just justice system.
All in all, Artificial Intelligence is just a tool like any other. In its perfect state it makes the world a better place. Yet with malicious human interference it can be used as a tool to influence societies decisions and it can also be used to control entire populations. In my opinion, nuclear power and AI are not that different. They are both tools that can influence entire populations for generations. So, it’s up to us to continuously regulate their use for good. For this to happen, people need to get educated on the threat that AI holds, because contrary to nuclear bombs, the negative impact it has is not tangible and will not be immediately felt by millions. The threat will creep into our society in a slow non-tangible manner.
Artificial Intelligence in itself should revolutionize our lives and should be of great positive impact to societies all over the world. But for this to happen we must stay vigilant and plan for the future in an ethical manner. We should never shy away from AI, we should embrace it, and make sure that it is being used for the good of humankind.