Opinion
In this article, I explore ‘bad’ AI; What can be the implications if we do not do something about it, how pervasive it is already in our society, and how can we save ourselves from it?
Published in · 6 min read · May 9, 2022
--
In 1963, Stanley Milgram did an experiment to understand what makes people obedient to authority. The experiment famously came to be known as the Milgram experiment.
The setup was the following:
The teacher (T) is asked by the experimenter (E) to provide an electric shock to the learner (L) every time the learner(L) gives a wrong answer. The electric shocks gradually go up to 750 volts, a shock that can even kill the learner (L). In reality, no electric shocks were given but the teacher (T) does not know that. The teacher (T) only hears sounds for each shock level, which was tape-recorded.
The experiment found that the majority of the teachers (T) ended up administrating the maximum shock to the learner (L) that could have killed the learner (L) if the shocks were given in real. Some of the experimenters protested giving the shocks during the experiment but nevertheless, continued the experiment.
This experiment showed that the majority of us will follow orders obediently from authority to the extent of even killing someone. We have experienced this time and again in history, people following orders to kill others, and when asked later why did they do so, their answers were ‘I was just following the order’.
What if the experimenter (E) in the Milgram experiment is an intelligent machine algorithm (or AI) that gives the orders instead of a human experimenter (E)? Maybe the teacher (T) does not even know that the experimenter is a robot. With the developments in speech generation, the algorithm does not even have to be smart at understanding the environment or context. It can be just a ‘dumb machine’ giving orders without knowing the context. Will people follow those orders? I think there is a real possibility that the majority of us will.
Is the scenario mentioned above a distant future, or something that we are already experiencing?
I am not one of the people who believe in a doomsday scenario when robots take over the world. However, we can see the trend already that AI algorithms are shaping our opinions and controlling the minds of the people.
Social media companies like Meta, Tiktok, and Twitter feed us with information and we start forming opinions. Scott Gallway in his article refers to social media as WMDs: weapons of mass distraction, whose goal is to divert attention away from physical life. For example, We can see that in the Ukraine-Russian war. As Putin loses an information war, he continues to kill thousands in a real war that we’re losing interest in.
We are getting so much into the virtual world that we are forgetting the real world and thus can be easily controlled, at scale, by algorithms that feed us with information or data through the virtual world.
The situation can get much worse: Given what we saw in the Milgram experiment, if some bad actors (or authority) decide to use the AI to control the minds of people at scale, then it is not that far when we follow orders to wage wars and kill others. We can already see a glimpse of this in social media where people fight, show hatred, and vilify each other without even knowing the other side well. Looking at the hate on social media, there is a real possibility that the ongoing Ukraine-Russia war can potentially lead us to World War 3.
There are two ways to control the growth of ‘bad’ AI: a) Through regulations, and b) we, the people, become aware and start doing something.
Regulations
Personally, I am never a big fan of regulations as it creates a top-down structure. Who will regulate the regulators? What if the regulators become bad actors? However, sometimes it is a necessary evil that we have to accept.
In recent times, we can see that politicians are catching up with what harm AI can do and are able to come up with good regulations. For example, the recent Digital Services Act (DSA) proposal from the EU will allow users to better understand how content is recommended to them and how moderation decisions are made. Users will also enjoy a right to reinstatement if platforms make mistakes. It might solve some issues but is not enough. There is always a chance that some rogue states and politicians will misuse regulations to control the freedom of people. The Milgram experiment already showed that we can easily give in to authority, so the dangers of ‘bad’ AI can be even worse if we solely rely on regulations. In addition, it takes significant time for the regulators to build a good regulation.
Collaboration, Compassion, and Consciousness (the 3 C’s)
The best bet we have to tackle ‘bad’ AI is we the people. We have to become more aware of what is happening and can happen, and build our own solutions that we want to use. Considering the tremendous impact that most of today’s AI solutions have on people and society, it would be anyway detrimental to build solutions in isolation from the people and social circ*mstances that make them necessary in the first place[1].
Collaboration among varied talents enables us to bridge gaps in understanding between different mindsets, share knowledge, and unite people and values. It, therefore, helps to create compassion and harnesses crowd wisdom, diversity, and inclusion to serve the long-term interests of those communities. However, it is important to not just have token diversity but true diversity of opinions.
The other key element is consciousness. As so much division exists in this world, we need to understand, that deep down we all are one, and all the groups that we associate with (nation, religion, sexual orientation, gender, the color of skin, etc) are just an illusion of our mind. We have the same fears, same aspirations, and similar dreams, and all our fate is the same. Thus our consciousness is collective.
If we want to survive and face future challenges of the world and from bad actors (and AI), we have to collaborate together with compassion and consciousness. Otherwise, we will make intelligent machines kill, fight and suppress each other. We can already see such AI algorithms suppressing the opinions of people who are different, built by big tech giants.
On the other hand, AI built by the three C’s (Collaboration, Compassion, and Consciousness) will help us to remove endemic sociological and historical bias and other inequalities that exist in society. AI does not challenge the status quo unless the process is purposely designed to do so.
“Never doubt that a small group of thoughtful, committed citizens can change the world.” — Margaret Mead
Right from the beginning of civilization, controlling people has always been the goal of the kings, and leaders. In past, through power and force, and now in the age of AI will be through data and knowledge by controlling our minds. This is very scary and with a click of a button and using algorithms, such control can be done much more effectively at scale.
But at the same time, never in the history of the earth, due to the access to knowledge being democratized so many people also have got the power — of the internet, connectivity, and knowledge. The same tools that bad actors can use, we the people also have access to and can use for good reasons. For the first time in history, so many people have got the power to make a change.
Machines are not that intelligent but, if we are not vigilant and aware enough, we will be controlled by bad actors. So it is our responsibility to build AI and future solutions with 3C’s (Collaboration, Compassion, and Consciousness) before it's too late.
Bio: I started Omdena, a collaborative platform to build AI solutions for the reasons mentioned in the article. Feel free to check the website and the work done.
Reference
In this comprehensive exploration of the implications of 'bad' AI, the author delves into the Milgram experiment from 1963, a seminal study on human obedience to authority. Drawing parallels between the experiment and the potential influence of AI algorithms, the author raises concerns about the current pervasiveness of AI in shaping opinions, particularly through social media platforms like Meta, TikTok, and Twitter.
The article outlines a scenario where AI algorithms, acting as authoritative figures, could manipulate individuals at scale, leading to potentially dire consequences such as increased social conflict and even the possibility of global conflicts like World War 3. The author emphasizes the need for intervention and proposes two approaches: regulations and collective awareness and action by individuals.
Regarding regulations, the author acknowledges their limitations, expressing skepticism about a top-down structure and the potential for misuse by rogue states and politicians. Despite recent efforts, such as the Digital Services Act (DSA) proposal from the EU, the author contends that regulations alone may not be sufficient to address the dangers posed by 'bad' AI.
The second approach, championed by the author, involves the collective efforts of individuals—referred to as the "three C's": Collaboration, Compassion, and Consciousness. The article stresses the importance of collaboration among diverse talents to bridge gaps in understanding, share knowledge, and unite people and values. Compassion is highlighted as a key element, emphasizing true diversity of opinions rather than token diversity. Consciousness, described as recognizing our collective identity beyond artificial divisions, is proposed as a means to foster unity and tackle challenges posed by 'bad' AI.
The author concludes by quoting Margaret Mead and asserting the power of a small group of committed citizens to change the world. They highlight the historical shift in power dynamics, with the democratization of knowledge and access to tools that both bad actors and individuals can leverage. The responsibility to build AI and future solutions with the three C's—Collaboration, Compassion, and Consciousness—is emphasized as a crucial step to prevent potential misuse and control by 'bad' actors. The article ends with a call to action and a reference to the author's initiative, Omdena, a collaborative platform for building AI solutions aligned with the proposed principles.
