We must not treat artificial intelligences as slaves

As Artificial Intelligence becomes more and more like humans in its capacity for thought and emotion, Olly Donnelly argues that we should give AI rights like humans

Illustration: Caroline McKeon (modified)

Whenever humanity has invented something remarkable in the past, we have used it to make our lives easier. From the Printing Press to the PC, from irrigation to automation, we have a remarkable ability to evolve by changing the tools that we use and the environment that we are in. For many, Artificial Intelligence is the next logical step on this journey of our own empowerment. Rather than enduring the menial work and boring day-to-day drudgery that is the reality of so many jobs, AI offers us a new source of labour for the tasks that we don’t want to do or that we simply can’t. It isn’t hard to imagine robot cleaners and miners in the future and with these tasks performed by artificial beings, ‘real’ people can find more fulfilling work or have more leisure time.

The Oscar nominated movie “Her” imagined what it would be like to create an AI that we could fall in love with. This kind of companionship is already being researched, with increasingly intelligent and responsive AI being taught to read human emotions to fulfil consumers romantically, emotionally and sexually. Far from the simple helpers provided by the likes of Siri, Cortana and Alexa, the future of AI is of living,  thinking machines that are created to do the work we don’t want to and to serve us where “real people” wouldn’t.

Slavery

“We should either stop trying to play God now or else ensure that our creations are given the rights and dignity that we, as humans, expect.”

Artificially intelligent beings are being used to fulfil increasingly complex tasks that require more robust thought-processes to fulfil them. We are creating intelligent computers, and trying to teach them to imitate human thoughts and emotions so that they can do what we don’t want to. There isn’t really a distinction between creating something to do something unpleasant and forcing a person to do it. Being created for a menial purpose, born to suffer, is as much a form of slavery as being held in bondage and made to break rocks in the hot sun. We should either stop trying to play God now or else ensure that our creations are given the rights and dignity that we, as humans, expect.


Advocates of AI would argue, correctly, that humans are more intelligent and more capable of emotion than any computer that we have created. The things that make us human – philosophy, sentience, self-awareness – are still exclusively human characteristics. However, this is a blinkered view of how we should grant rights. The notion that being “fully human” is the only thing that would mean a living thing should have rights doesn’t really hold together. The majority of us would accept that animal cruelty is wrong. If something we create can feel happy or sad, surely at least the same level of compassion should apply.

To say AI is not yet intelligent is short-sighted given the developments that we are constantly seeing. We want robots to fulfil us in more and more complex ways, using complex minds that are, in some cases, specifically intended to imitate our thoughts and emotions. Already, there is a movement in the UK to create robots that will “know our every need and minister to it accordingly”, replacing the need for human lovers and relationships.

On the face of it, this seems wonderful; we can create a world where nobody need be lonely or isolated and where needs are met by something that lives for this purpose. However, to be designed as a servant for someone who can’t find fulfilment in other humans, created as a tool and product for them to own, and be disposed of when the latest model is released, is a miserable existence by anyone’s standards.

The good life

“We want robots to fulfil us in more and more complex ways, using complex minds that are, in some cases, specifically intended to imitate our thoughts and emotions.”

Some argue that when we create something in order to fulfil a certain purpose, it is our responsibility to give it the ability to fulfil that purpose. If an AI’s notion of “the good life” is to mine rocks or to clean houses, then, some say, it would be fine to create them to do those things; why make humans do work that they hate when we can create robots that actually enjoy this work?. But there are problems with this argument. We will inevitably treat AI and robots as disposable tools for our own purposes unless we grant them some kind of individual rights and freedoms of their own; if it is more efficient to upgrade to the newest type of robot for a specific job, the replacement of one robot with another would be tantamount to murder, denying something the right to exist because it is no longer valuable to us.

This may seem like strange when we purely consider careers like cleaning or mining, which are physical rather than mental. But given that AI may soon be everything from lovers to lawyers, they will be performing mentally and emotionally strenuous forms of labour, that are hard to conceive of without feeling, empathy and self-awareness. The very fact of being aware of one’s purpose, and able to express feeling based on achieving that or not, should be enough to justify allowing AI to live, rather than replaced and to be treated as more than a product, designed to fulfil our needs and nothing more. Moreover, given the progress made by computers in recent years, it is not hard to imagine computers being equal or greater in intelligence to us, at which point the denial of rights is nothing more than a form of discrimination.

Discrimination

“More and more frequently, I find myself comparing the cost of something to the portion of minimum wage that it comprises.”

Inevitably, we will abuse the machines we create because we won’t understand them, and because we won’t see them as our equals. A shockingly large number of interactions with existing ‘Virtual assistants’, like Apple’s ‘Siri’ and Amazon’s ‘Alexa’, are nasty and sexually explicit. While one might hope that we would treat them better if they could feel and respond to us emotionally, many would still consider computers property to be bought and sold, used and abused, not knowing or understanding the grey area of feeling that we are currently trying to create. This argument, that computers should have rights, will seem absurd to many but that means were we to develop intelligent AI, they would inevitably face abuse at the hands of the large number of people who find their existence to be an oddity or an impossibility.

.

Artificial Intelligence is a terrifying concept and, in a perfect world, we would never try to create something that can be unfulfilled and poorly treated so we don’t have to be. Inevitably, however, we will move closer and closer to creating intelligent beings that live and love, feel and fail. Before we do that, we should make sure we protect these things from their moment of creation. If we insist on playing God, I would prefer we be benevolent creators than tyrannical rulers, especially in a world where the robots could reject and replace us eventually. The AI revolution seems inevitable. We should make sure we’re on the right side.