I sometimes think that, although there is currently a great inequality in the world, there will come a time when we will have found ways of making society more equal and particular groups in society will no longer be disadvantaged. We are living at a time of almost unprecedented skepticism and upheaval of both our systems of government and economics. At the same time, over the next couple of decades artificial intelligence will have advanced to such a degree that it will be necessary to build some kind of morality into all artifacts that do any kind of complex calculation that has significance for safety or wellbeing. These two ideas are connected. They both require a more sophisticated approach to the way in which we allocate moral responsibility.
Perhaps this is pie in the sky, because all we have achieved since the millennium are greater degrees of inequality, and although autonomous vehicles are now on the horizon we have still not managed to get computers or robots to perform the kind of tasks that are quintessentially human like having compassion or pouring a cup of tea. Nevertheless stay with me in this thought experiment and see where it goes. I’m writing this mainly from a UK perspective but the points apply quite generally.
Assume, just for one moment, that in the next 10 to 20 years, in some countries at least, automation has encroached on not just most manual jobs but also on more highly skilled and professional work. There are many commentators that support this view. Assume also that the product of this is distributed fairly (a far less likely assumption) such that everybody in society has more than enough to meet their basic needs. What’s left to do?
One possibility is that just like on a Christmas Day when we are relieved of the normal routine of work and thrown together with those closest to us, one reaction will be to pick holes in each other’s behaviour. Another will be a spate of litigation against those organisations that have created artificially intelligent machines that for one reason or another go wrong, cause harm or even just frustrate and annoy. In other words, we will all become involved in the process of ‘debugging’ the unacceptable and often unintended consequences of the artifacts and systems we build (including, by a stretch of the imagination, all our interpersonal relationships and contractual arrangements – implicit and explicit).
The identification of the cause of something undesirable is to allocate blame (if its desirable we praise). We commonly engage in blame when haven’t got anything more pressing to do. So, when our resources are freed up from the need to put food on the table or engage in all that other spurious activity that society and the commercial world would have us believe is important, our attention can turn to fixing those things that are bugging us.
This is probably something that we are better at doing the older we get because it takes some experience and maturity to distinguish between the things that we ourselves cause and the things that are caused by others. A baby has little sense of self. Teenagers must, sometime agonizingly, go through the process of establishing their own identity. Even as adults we often have difficulty working out whether, say, Microsoft is at fault or whether it is our own stupidity (to my mind it makes more sense to always blame Microsoft!).
In other words, I can see the theme of blame manifesting itself more strongly in all sorts of systems, both formal and informal, and partly through want of anything better to do, an industry of blame related services will mushroom. These are services like mediation, arbitration and litigation. We believe these activities are driven by human values, and therefore ultimately rely on human skill and judgment. However, in a complex world in which it is possible to go on forever establishing cause, culpability and moral responsibility, what we have now by way of legal argument will start to look like child’s-play. As we continue to make progress in both our scientific and philosophical understandings of the nature of intention and free will, we will be far better equipped to make detailed and complex arguments about who or what is to blame for anything that we are not happy with.
So, for example, if we have each hired autonomous vehicles and your hired vehicle runs into my hired vehicle maiming one of my passengers, we enter a world in which the attribution of blame might have to be shared between the manufacturers of the vehicles, their owners, their hirers, the authorities that maintain the road, other people and autonomous agents that might have contributed to the accident, the passengers of the vehicles and their insurers and so on.
From what I’ve seen of the way the car insurance industry works at the moment it would seem to be based on a set of clandestine rules and arrangements for settlement between insurers that the general public is rarely aware of. These are no doubt to preserve the profitability of the insurance industry and it is difficult to tell whether they serve the interests of the insured. It would therefore be no bad thing if the logic on which blame is allocated, and settlements were made, were opened to public scrutiny.
At the other end of the spectrum, if person A says something to person B that makes them unhappy, can B legitimately blame A? Perhaps they can in the sense of the superficial cause, but whether they can in the sense of a moral responsibility is more difficult to establish and may depend on unpacking the nature of the relationship between A and B. This is to examine the implicit contract between A and B and to see whether the comment falls, inside or outside the current ‘rules of the game’ or indeed whether it represents a legitimate attempt to shift the rules. If A hurts B by saying that C does not like them, this might be morally praiseworthy within the set of rules we define as ‘friendship’, especially if it’s true and it helps B understand C’s behaviour.
Blame is a complex concept. It involves notions of causality, intention, consequence, and the evaluation of moral responsibility to say nothing of the roles of compassion and mitigation. As our understanding of the brain, psychology and society increases, the more difficult it becomes to identify a single point of blame. Is the person with a brain injury responsible for their actions? Can a child coming from a family of many generations of abuse and social disadvantage/injustice, brought up on computer games and films portraying violence, and exposed to extensive media coverage of terrorist acts, provoked and pressurized by adverse circumstances, to be held solely responsible for their antisocial behaviour? The more we discover, the more we appreciate that we can only understand individual behaviours as taking place within a socially constructed context. They are manifestations of entire systems as opposed to individually calculated and motivated.
The law assesses culpability in terms of intent. An act that causes harm (or a failure to act in foreseeable circumstances) can be categorized as purposeful, reckless or negligent. The law can also take into account consequence, especially when considering the victim, and how the extent of punishment will affect them. However the legal system has many anomalies and drawbacks. It’s only virtue is that it is a more or less agreed system of reconciling differences of opinion (on what can and cannot be done in a criminal or civil sense). This does not mean that it is reasonable or fair.
Our legal system as it stands, favours those who can construct the best argument. It is a mechanism that allocates culpability and blame based on the capability to research and articulate the argument, or pay for this to be done. This further disadvantages those that are already disadvantaged.
One way out of all this is to put aside notions of morality and ethics and look at blame simply as a process of fixing a problem. Although our emotions of frustration and anger drive us towards punishment and retribution, if we can put this aside, it is possible to establish a much cleaner logic for the allocation of blame and see blame , not as a justification for retribution, but as a mechanism of identifying what exactly needs to be fixed. This does not mean giving up the evaluation of behaviour. We can still say that a particular act is ‘good’ or ‘bad’ in terms of it’s intention and it’s outcome, but instead of seeking punishment and retribution, focus on ‘fixing the problem’.
Punishment may fix the problem in terms of acting as a deterrent. Incarceration may fix the problem in the sense of removing the problem from mainstream society. And these, therefore act as legitimate justifications. However, these mechanisms have some unfortunate unintended consequences. They stigmatise and further disadvantage the perpetrator and they put the perpetrator into a culture where unsocial acts are the norm and new criminal techniques can be easily learnt. In many ways they do precisely the opposite of what is required to solve the problem. When, as a society are we going to grow up and address this head-on.
Fortunately, there are some new approaches, like restorative justice, where the perpetrator and the victim are brought together, appear to work to the advantage of both. There will always be occasions where removal from society is the only safe option, but in most cases of crime, this is just sweeping the problem under the carpet. The number of people in prison that have mental health issues is a travesty and a testament to just how bad a system we have built. Particularly in the US. See:
So, where is all this going? We can see that:
- Our systems of justice are often unfair and can aggravate rather than fix the problem. They, themselves need fixing.
- That there will be an increasing need (and hopefully capacity) to grow service industries in all the activities that deal with relationships between people, and between people and society – services such as mediation, counseling, arbitration and more productive forms of litigation and redress.
- That as our artifacts grow in the capacity to act autonomously, there will be an increasing need to embed into them, explicit values and some kind of moral ‘code’
- It is becoming increasingly apparent that we need a systemic approach to understanding individual behaviours
- That advances in science, psychology, neurology, philosophy and computer science offer more informed and more effective ways of doing this.
We should be working on new, more explicit, evidenced based, and rigourous notions of justice, responsibly and blame that do not perpetuate the faults of our current systems.
As our technologies and understandings develop, we are going to need new regulators of a society, containing both human and artificial agents, each constructing and executing a complex web of intentions within a new social context. If we get this right, there is a growth industry of people based services that can lead the way to a more productive and fairer society.