Facebook’s ambition into AI: What are the philosophical implications?

LISA EADICICCO (2019) from Business Insider has reported that Facebook’s AI (Artificial Intelligence) division is researching and investing into AI that can learn new skills without prior information, learn new skills quickly through creating a ‘curiosity’ algorithm, and allowing the AI to respond off stimuli through human senses such as ‘touch’. Eadicicco also mentions that Google and Apple are investing in their own AI research, and that this area is a booming area of investment.

What would succeeding in these goals tell us about self-aware, or conscious AI?

Alan Turing’s ‘Turing Test’ says that we can say that a machine can be said to be thinking just like human beings, when a human made to have a conversation with an AI (but does not know it’s one) but cannot tell it apart from a human. If facebook were indeed able to learn in such a way, they would certainly be on the path to passing the Turing Test. However, the Turing Test has been challenged through compelling thought experiments. One is John Searle’s ‘Chinese room’ thought experiment, where a person in a separate room, who cannot understand Chinese, is sliding Chinese messages to the Chinese person under the door. These messages have been coded to help the person to respond appropriately to the Chinese person. So, the Chinese person thinks they are recieving messages from someone who understands Chinese, when they really don’t.

Hence, the Turing Test does not tell us that if AI can behave just like humans, then they are experiencing self-awareness or are conscious. They could be just behaving as if they are through their coding and not experience any kind of self-awareness or inner life. So, such inventions by Facebook would be very impressive, but would not (at least yet) be giving us artificial human beings.

Economic implications: Jobs

One of the most immediate concerns of AI, is not a fear of Terminator situations of robots enslaving humans, but a fear of losing jobs. Whenever a computer or a machine can be taught to do a task a human does, it will almost always do it better than human will. Once you program a computer how to play Chess, it will always beat the world’s best Chessmaster. So, when a business can get machines to do a human’s job, it will automate that job since it would be cheaper and more efficient to do so.

The fear of automation has brought back a relatively old idea of a Universal Basic Income (UBI), which is a no questions asked salary to every citizen from the government, whether you are working or you are not. The most compelling arguments for a UBI come from philosopher Phillipe Van Parjis, which can be read here: https://bostonreview.net/archives/BR25.5/vanparijs.html

This specific AI Facebook is able to be done, then this AI could very well not only be able to do tasks required of it, but also be able to express creativity, to be able to come up with new ideas. This would indeed threaten one of the main advantages human beings have over machines. Even when a machine can do what is instructed better than a human being, it is us human beings that come up with better and different ways to do things. So, success in creative AI would further the fear of automating human beings out of usefulness in the workplace.

Private and State ownership: AI a new ‘Arms race’?

Given the competition between Facebook, Google, and Apple on who will be the first to cross the line on creative, self-learning AI, this may raise the question of whether one company should own the Patents to such technology? If the reliance of such technology became universal, such as all companies around the world depended on it for running their businesses, then how could any other company would be able to compete? Furthermore, one company, such as Facebook, holding all the economic power, would inevitably hold political power, similar to banks that are ‘too big to fail’.

One answer could be for the state to intervene and allow pubilc ownership of the technology. However, if Facebook, Apple, and Google were to do this first, and either decided not patent it or allow it to be State/Publicly owned (I do concede this situation is fanciful), one country would hold the technology. This power could be used to create more sophisticated drones and other military technology. Such power, could possibly be analgous to one country holding all of the world’s nuclear warheads. So, a possible option would be to allow this technology to be owned globally, not through one company or one country.

Not all doom and gloom

The positive outcomes from such technology, is the ability for machines to undertake most of our labour, either in the workplace or in our day to day lives. This will leave us available to persue more of our own personal interests. This is not advocating for a hedonistic lifestyle. We would be more free to create Art, socialise, learn history, engage more deeply with our own communities and other communities, and so on. Some may retort that our identity is tied to what we do for a living, and without this, we would lose our sense of meaning. Of course in a capitalist society this is true right now, but as the world changes, and we move away from only looking at our place in the world through an economic lens, our sense of meaning will change as well.

Reference

https://www.businessinsider.com.au/facebook-secretive-ai-robotics-division-revealed-2019-5?r=US&IR=T&fbclid=IwAR0pu8QWq96oE-WVLvzqSundc7nIfgN1VnxI6Z_AKXyx5k06cSU6ppg82K0