When AI is mentioned it is only a matter of time before the word ‘bias’ is heard. They seem to go together like ping and pong, especially in debates around AI in education. Yet the discussions are often merely examples of bias themselves – confirmation, negativity and availability baises. There’s little analysis behind the claims. ‘AI programmers are largely white males so all algorithms are biased - patriarchal and racist’ or the commonly uttered phrase ‘All algorithms are biased’. In practice, you hear the same few examples being brought up time and time again: black face/gorilla, recruitment and reoffender software. Most of these examples have their origin in Cathy O’Neil’s Weapons of Math destruction orminternet memes. More of this later.
To be fair AI is for most an invisible force, the part of the iceberg that lies below the surface. AI is many things, can be opaque technically and true causality difficult to trace. So, to unpack this issue it may be wise to look at the premises of the argument, as this is where many of the misconceptions arise.
Coders and AI
First up, the charge that the root cause is male, white coders. AI programmers these days are more likely to be Chinese or Indian than white. AI is a global phenomenon, not confined to the western world.
The Chinese government has invested a great deal in these skills through Artificial Intelligence 2.0. The 13th Five-Year Plan (2016-2020), the Made in China 2025 program, Robotics Industry Development Plan and Three-Year Guidance for Internet Plus Artificial Intelligence Plan (2016-2018) are all contributing to boosting AI skills, research and development. India has an education system that sees ‘engineering’ and ‘programming’ as admirable careers and a huge outsourcing software industry with a $150 billion IT export business.
Even in Silicon Valley the presence of Asian and Indian programmers is so prevalent that they feature in every sitcom on the subject. Even if the numbers are wrong the idea that coders infect AI with racist code, like the spread of Ebola, is ridiculous. One wouldn’t deny the probable presence of some bias but the idea that it is omnipresent is ridiculous.
Gender and AI
True there is a gender differential, and this will continue, as there are gender differences when it comes to focused, attention to detail coding in the higher echelons of AI programming. We know that there is a genetic cause of autism, a constellation (not spectrum), of cognitive traits and that this is weighted towards males (and no it is not merely a function of underdiagnosis in girls). For this reason alone there is likely to be a gender difference in high-performance coding teams for the foreseeable future. In addition, the idea that these coders are unconsciously, or worse, consciously creating racist and sexist algorithms is an exaggeration. One has to work quite hard to do this and to suggest that ALL algorithms are written in this way is another exaggeration. Some may, but most are not.
Anthropomorphic bias and AI
The term Artificial Intelligence can in itself be a problem, as the word ‘intelligence’ is a genuinely misleading, anthropomorphic term. AI is not cognitive in any meaningful sense, not conscious and not intelligent other than in the sense that it can perform some very specific tasks well. It may win at chess and GO but it doesn’t know that it even playing these things, never mind the fact that it has won.
Anthropomorphic bias appears to arise from our natural ability to read the minds of others and therefore attribute qualities to computers and software that are not actually there. Behind this basic confusion is the idea that AI is one thing – it is not – it encapsulates 2500 years of mathematics since Euclid put the first algorithm down on papyrus and there are many schools of AI that take radically different approaches. The field is an array of different techniques, often mathematically, quite separate from each other.
ALL humans are biased
First, it is true that ALL humans are biased, as shown by Nobel Prize winning psychologist Daniel Kahneman and his colleague Amos Tversky, who exposed a whole pantheon of biases that we are largely born with and are difficult to shift, even through education and training. Teaching is soaked in bias. There is socio-economic bias in policy as it is often made by those who favour a certain type of education. Education can be bought privately introducing inequalities. Gender, race and socio-economic bias is often found in the act of teaching itself. We know that gender bias is present in subtly directing girls away from STEM subjects and we know that children from lower socio-economic groups are treated differently. Even, so-called objective assessment is biased, often influenced by all sorts of cognitive factors – content bias, context bias, marking bias and so on.
Bias in thinking about AI
There are several human biases behind our thinking about AI.
We have already mentioned Anthropomorphic bias, where reading ‘bias’ into software is often the result of this over-anthropomorphising.
Availability bias arises when we frame thoughts on what is available, rather than pure reason. So crude images of robots enter the mind as characterising AI, as opposed to software or mathematics, which is not, for most, easy to call to mind or visualise. This skews our view of what AI is and its dangers, often producing dystopian ‘Hollywood’ perspectives, rather than objective judgement.
Then there’s Negativity bias, where the negative has more impact than the positive, so the Rise of the Robots and other dystopian visions come to mind more readily than positive examples such as fraud detection or cancer diagnosis.
Most of all we have Confirmation bias, that leaps into action whenever we hear of something that seems like a threat and we want to confirm our view of it as ethically wrong.
Indeed, the accusation that all algorithms are biased is often (not always) a combination of ignorance about what algorithms are and a combination of four human biases – anthropomorphism, availability, negativity, confirmation and anthropomorphism bias. It is often a sign of bias in the objector, who wants to confirm their own deficit-based weltanschauung and apply a universal, dystopian interpretation to AI with a healthy dose of neophobia (fear of the new).
ALL AI is not biased
You are likely in your first lesson on algorithms to be taught some sorting mechanisms (there are many). Now it is difficult to see how sorting a set of random numbers into ascending order can be either sexist or racist. The point is that most algorithms are benign, doing a mechanical job and free from bias. They can improve performance in terms of strength, precision and performance over time (robots in factories), compressing and decompressing comms, encryption algorithms, computational strategies in games (chess, GO, Poker and so on), diagnosis-investigation-treatment in healthcare and reduced fraud in finance. Most algorithms, embedded in most contexts are benign and free from bias.
Note that I said ‘most’ not ‘all’. It is not true to say that all algorithms and/or data sets are biased, unless one resorts to the idea that everything is socially constructed and therefore subject to bias. As Popper showed, this is an all-embracing theory to which there is no possible objection, as even the objections are interpreted as being part of the problem. This is, in effect, a sociological dead-end.
Bias in statistics and maths
Al is is not conscious or aware of its purpose. It is, as Roger Schank kept saying, just software, and as such, is not ‘biased’ in the way we attribute that word to ‘humans’. The biases in humans have evolved over millions of years with additional cultural input. AI is maths and we must be careful about anthropomorphising the problem. There is a definition of ‘bias’ in statistics, which is not a pejorative term, but precisely defined as the difference between an expected value and the true value of a parameter. If the value is zero, it is called unbiased. This is not so much bias as a precise recognition of differentials.
However, human bias can be translated into other forms of statistical or mathematical bias. One must now distinguish between algorithms and data. There is no exact mathematical definition of ‘algorithm’ where bias is most likely to be introduced through weightings and techniques used. Data is where most of the problems arise. One example is poor sampling; too small a sample, under-representations or over-representations. Data collection can also have bias due to faulty data gathering in the instruments themselves. Selection bias in data occurs when it is gathered selectively and not randomly.
However, the statistical approach at least recognises these biases and adopts scientific and mathematical methods to try to eliminate these biases. This is a key point – human bias often goes unchecked, statistical and mathematical bias is subjected to rigorous checks. That is not to say that it is flawless but error rates and attempts to quantify statistical and mathematical bias have been developed over a long time, to counter human bias. That is the essence of the scientific method.
The word ‘algorithm’ induces a sort of simplistic interpretation of AI. Some algorithms are not created by humans, code can create code, some are deliberately generated in evolutionary AI to create variation and then selection against a fitness purpose. It’s complex. There are algorithms in nature that determine genetic outcomes, the way plants grow and many other natural phenomena. Some thing that there is a set of deep algorithms that determine the whole of life itself. Evolutionary AI allows algorithms to be promulgated or generated by algorithms themselves, in an attempt to mimic evolution, but defining fitness and selecting those that work. While it is true that bias can creep into this process it is wrong to claim that all algorithms are created solely by the hand of the coder.
AI and transparency
A common observation in contemporary AI is that its inner workings are opaque, especially machine learning using neural networks. But compare this to another social good – medicine. We know it works but we don’t know how. As Jon Clardy, a professor of biological chemistry and molecular pharmacology at Harvard Medical School says, "the idea that drugs are the result of a clean, logical search for molecules that work is a ‘fairytale'”. Many drugs work but we have no idea why they work. Medicine tends to throw possible solutions at problems, then observe if it works or not. Now most AI is not like this but some is. We need to be careful about bias but in many cases, especially in education, we are more interested in outputs and attainment, which can be measured in relation to social equality and equality of opportunity. We have a far greater chance of tackling these problems using AI than by sticking to good, old-fashioned bias in human teaching.
Fail means First Attempt In Learning
Nass and Reeves through 35 studies in The Media Equation showed that the temptation to anthropomorphise technology is always there. We must resist the temptation to think this is anything but bias. When an algorithm, for example, correlates a black face with a gorilla, it is not that it is biased in the human sense of being a racist, namely a racist agent. The AI knows nothing of itself, it is just software. Indeed, it is merely an attempt to execute code and this sort of error is often how machine learning actually learns. Indeed, this repeated attempt at statistical optimisation lies at the very heart of what AI is. Failure is what makes it tick. The good news is that repeated failure results in improvement in machine learning, reinforcement learning, adversarial techniques and so on. It is often absolutely necessary to learn from mistakes to make progress. We need to applaud failure, not jump on the bias bandwagon.
When Google was found to stick the label of gorilla on black faces in 2015, there is no doubt that it was racist in the sense of causing offence. Rather then someone being racist in Google, or having a piece of maths that is racist in any intentional sense, this is a systems failure. The problem was spotted and Google responded within the hour. We need to recognise that technology is rarely foolproof, neither are humans. Failures will occur. Machines do not have the cognitive checks and balances that humans have on such cultural issues but they can be changed and improved to avoid them. We need to see this as a process and not just block progress on the back of outliers. We need to accept that these are mistakes and learn from these mistakes. If mistakes are made, call them out, eliminate the errors and move on. FAIL in this case means First Attempt In Learning. The correct response is not to define and dismiss AI because of these failures but see them as opportunities for success.
The main problem here, is not the very real issue of emanating bias from software, which is what we must strive to do but the simple contrarianism behind much of the debate. This was largely fuelled by one book….
Weapons of 'Math' Destruction - sexed up dossier on AI?
Unfortunate title, as O’Neil’s supposed WMDs are as bad as Saddam Hussein’s mythical WMDs, the evidence similarly weak, sexed up and cherry picked. This is the go-to book for those who want to stick it to AI by reading a pot-boiler. But rather than taking an honest look at the subject, O’Neil takes the ‘Weapons of Math Destruction’ line far too literally, and unwittingly re-uses a term that has come to mean exaggeration and untruths. The book has some good case studies and passages but the search for truth is lost as she tries too hard to be a clickbait contrarian.
The first example borders on the bizarre. It concerns a teacher who is supposedly sacked because an algorithm said she should be sacked. Yet the true cause, as revealed by O’Neil, are other teachers who have cheated on behalf of their students in tests. Interestingly, they were caught through statistical checking, as too many erasures were found on the test sheets. That’s more man than machine.
The second is even worse. Nobody really thinks that US College Rankings are algorithmic in any serious sense. The ranking models are quite simply statistically wrong. The problem is not the existence of fictional WMDs but poor schoolboy errors in the basic maths. It is a straw man, as they use subjective surveys and proxies and everybody knows they are gamed. Malcolm Gladwell did a much better job in exposing them as self-fulfilling exercises in marketing. In fact. most of the problems uncovered in the book, if one does a deeper analysis, are human.
Take PredPol, the predictive policing software. Sure it has its glitches but the advantages vastly outweigh the disadvantages and the system, and its use, evolve over time to eliminate the problems. The main problem here is a form of bias or one-sidedness in the analysis. Most technology has a downside. We drive cars, despite the fact that well over a million people die gruesome and painful deaths every year from in car accidents. Rather than tease out the complexity, even comparing upsides with downsides, we are given over-simplifications. The proposition that all algorithms are biased is as foolish as the idea that all algorithms are free from bias. This is a complex area that needs careful thought and the real truth lies, as usual, somewhere in-between. Technology often has this cost-benefit feature. To focus on just one side is quite simply a mathematical distortion.
The chapter headings are also a dead giveaway - Bomb Parts, Shell Shocked, Arms Race, Civilian Casualties, Ineligible to serve, Sweating Bullets, Collateral Damage, No Safe Zone, The Targeted Civilian and Propaganda Machine. This is not 9/11 and the language of WMDs is hyperbolic - verging on propaganda itself.
At times O’Neil makes good points on ‘data' – small data sets, subjective survey data and proxies – but this is nothing new and features in any 101 statistics course. The mistake is to pin the bad data problem on algorithms and AI – that’s often a misattribution. Time and time again we get straw men in online advertising, personality tests, credit scoring, recruitment, insurance, social media. Sure problems exist but posing marginal errors as a global threat is a tactic that may sell books but is hardly objective. In this sense, O'Neil plays the very game she professes to despise - bias and exaggeration.
The final chapter is where it all goes badly wrong, with the laughable Hippocratic Oath. Here’s the first line in her imagined oath “I will remember that I didn’t make the world, and it doesn’t satisfy my equations” a flimsy line. There is, however one interesting idea – that AI be used to police itself. A number of people are working on this and it is a good example of seeing technology realistically, as being a force for both good and bad, and that the good will triumph if we use it for human good.
This book relentlessly lays the blame at the door of AI for all kinds of injustices, but mostly it exaggerates or fails to identify the real, root causes. The book is readable, as it is lightly autobiographical, and does pose the right questions about the dangers inherent in these technologies. Unfortunately it provides exaggerated analyses and rarely the right answers. Let us remember that Weapons of Mass Destruction turned out to be lies, used to promote a disastrous war. They were sexed up through dodgy dossiers. So it is with this populist paperback.
This is an important issue being clouded by often uninformed and exaggerated. Positions. AI is unique, in my view, in having a large number of well-funded entities, set up to research and advise on the ethical issues around AI. They are doing a good job in surfacing issues, suggesting solutions and will influence regulation and policy. Hyperbolic statements based on a few flawed meme-like cases do not solve the problems that will inevitably arise. Technology is almost always a balance up upsides and downsides, let’s not throw the opportunities in education away on the basis of bias, whether in commentators or AI.