Ever since the launch of ChatGPT at the end of 2022, everyone seems to have something to say about artificial intelligence. Due to several dystopian films which made a lasting impression of AI destroying humanity, most notably the Terminator franchise, there is widespread fear and conspiracies on the internet of how Skynet (the fictional AI network in Terminator) could come to be. Spoiler alert: while there are many concerns about AI worth worrying about, Skynet is NOT it. There is so much misinformation and misunderstanding about the whole ordeal that I think most people are in need of some clarity.
For that, I strongly recommend everyone to listen to the 2021 BBC Reith Lecture series by British computer scientist Professor Stuart Russell on what could be the biggest event in human history, the emergence of general purpose artificial intelligence. In this article, I wish to discuss some of the issues which are relevant to us right now from the series of lectures.
A type of artificial intelligence which we are often subjected to is the social media content-selection algorithms, according to Professor Russell, not particularly intelligent, but have "more power over people’s cognitive intake than any dictator in history". The algorithm's objective is to maximise click-through, the idea is to have the algorithm send posts to the user which they will like, but to achieve its objective, the algorithm learns to modify the state of the user's mind, making the user more predictable, and so more likely to click on posts given to them. Users with extreme views are more predictable, the algorithm preys on that, leading to the growth of extremism all around the world. Professor Russell suggests that instead of giving specific objectives to the AI (the machine), "the machine’s only objective is to maximise the realisation of human preferences," to further understand this idea, please listen to the fourth installment of his lecture series.
Moving on to a physical threat, lethal autonomous weapon systems, that is according to the United Nations, “weapons that locate, select, and engage human targets without human supervision," "engage", an euphemism for "kill". Images of Terminator robots often accompany reports on autonomous weapons. This is misleading because for one, this gives off the impression that they are fictional when in fact autonomous weapons exist and are on sale right now on the web, also, it suggests that the problem is Skynet, the emergence of consciousness in AI leading to its hatred for us humans when in fact Skynet was never the problem.
The problem of autonomous weapons is when they are programmed to kill specific humans, those humans possibly including you and your loved ones. If you think no one would program a machine to kill civilians, then I am afraid you are truly naive in your beliefs, especially when you consider the current conflicts in different parts of the world. There is also the risk of autonomous weapons being subjected to cyber infiltration, leading to them turning on their owners during a war; and the risk of escalation when the system overreacts or retaliates to a false alarm, escalating a conflict into war.
The US Deputy Secretary of Defense in 2016 told Professor Russell that they have carefully listened to the arguments and his experts assured him that they would not accidentally create Skynet after calls for autonomous weapon ban, clearly, the message had not gotten through.
In the Convention for Certain Conventional Weapons (CCW), only weapons which operate “completely outside any framework of human control” are banned, so basically they banned Skynet. Current US policy states that "Autonomous … weapons systems shall be designed to allow commanders and operators to exercise appropriate levels of human judgment over the use of force." The UK Ministry of Defence stated that "there must be context-appropriate human involvement in weapons which identify, select and attack targets" and that the UK "does not possess fully autonomous weapon systems and has no intention of developing them".
As we head towards a future of artificial intelligence, we may not all have to understand the inner workings of an AI, we need to understand the effects it would have on our lives, how to prevent/stop dire consequences, how to use AI to our society's benefits, and most importantly, how to live with it.
Learn more from The Reith Lectures: The Biggest Event in Human History on BBC Sounds
