Human Compatible by Stuart Russell - go to guy on AI - a must read..

Unlike Bostrum’s Superintelligence, this book is eminently readable. Russell is not only an AI guy, he is a renowned educator and has won awards as an educator. Along with Peter Norvik, he wrote the classic, globally recognised textbook AI Artificial Intelligence: A Modern Approach and has been in AI long before most of the arrivistes currently bashing out books on AI and ethics.

He anchors his ethical arguments on the problem of ‘control’, machines that achieve our not their objectives. Unlike many of the popular books on ethics in AI, from contrarians like Cathy O’Neil, Safiya Noble and Shoshana Zuboff, who tend to throw the babies and bathwater out with the bath, he is level-headed and draws upon his decades of experience as a pioneer in the field to explain the flaws, limitations and dangers, as well as the power, promise and potential of AI.

He takes on his opposition directly and unlike Kevin Kelly, Andrew Ng and Stephen Pinker, he thinks that Superintelligence is a possibility. This is a brilliant summary of the main positions with Musk, Hawking and Gates at one end of a spectrum and Pinker at the other. Russell is in the middle, scared but offering a change of direction and solutions.

AI and learning
He points towards significant advances through AI in health, finance and education. The Global Learning XPRIZE is cited (you never hear about this but it’s important) and gives us 80 years before human-level AI arrives. Note that super-competent teaching AI may arrive decades before this. Understanding language is essential as is understanding and building hierarchies of abstraction. He scotches the idea of robots carrying around their own brains (therefore robot teachers) in favour of connected computing, EaaS, Everything as a Service. The low hanging fruit may well be in accelerating learning in humans. AI tutors, he thinks, will deliver learning to all, at negligible cost.

Unlike most treatments of ‘bias in AI’ Russell remains cool and objective in surfacing the problems but also patiently explaining the solutions. He avoids the usual teeth gnashing about capitalism, ethicswashing and ethicsbashing, for a typically balanced analysis of the problems but also the solutions – pre-and post processing, embedding fairness criteria and so on. 

Beneficial AI
Provably beneficial AI systems will need clear principles around altruism, humility and reference to human preferences. Beneficial AI is AI where we know what the purpose is. This is not as easy as it sounds and he uses AlphaGo to show how tricky it is -  is the goal 'winning' or 'reinforcement learning where winning is the reward'?

Preference Utilitarianism
The complication in all this is we humans. His take on moral philosophy is utilitarian, or rather Harsanyi’s preference utilitarianism where AIs only purpose is to manifest human preferences. His defence of this position is impressive and if you want to have a position on AI and ethics this is well worth reading as he rightly takes ethics (moral philosophy) seriously – many don’t which leads to fantasy castles of certainty built on quicksand.

Everyone and their uncle
The old model of technology that meets an objective is no longer valid. We need technology that defers to humans and can be switched off. Rather critical that ‘everyone and their uncle’ is setting up an AI and ethics board, department, framework etc., he proposes we give the top institutions a chance here – AAAI, IEEE, Partnership on AI and EU. It is not that others do not matter, only that there is too much low-level noise.

The strength of the book is that it is neither dystopian nor utopian. His line is realism, both about the dangers and opportunities. This is a book of some depth as he faces up to the possibility of an existential threat not with platitudes, anthropomorphism or manifestos butvreal solution in terms of moral philosophy and practical steps.

Of all the books I’ve read on AI and ethics, this, along with pieces from Dennett and Pinker have been the best. These three have truly shaped my own opinions on ethics in AI. I lean perhaps more to Dennett and Pinker in thinking the problem of superintelligence is not as likely, extreme or unmanageable as many suggest but this book is as good as you get on facing up to the perils, even if they are improbable. AI poses risks he but remains optimistic that we can retain control and harness the technology for good. 

AI and ethics - it's not as good as you think and not as bad as you fear

Joanna Bryson, one of the world’s experts in AI and ethics is right when she points out that the big problem in AI and Ethics is ‘anthropomorphising’. AI is competence without comprehension. It can beat you at chess, Go and poker but doesn’t know it has won. Literally hundreds of AI and ethics groups have sprung up over the last couple of years. Some are serious international bodies like the EU, IEEE and so on, but it is important to examine but remain level-headed on this issue. The danger is that we destroy the social goods that AI offer, by demonising it  before it has been tried.

Having just launched a new book ‘AI for Learning’ in which I tackle these ethical issues in some detail, I thought I’d provide a taster for the ethical concerns as they may affect the world of learning. 

Existential

Let’s get one moral issue out of the way – the existential threat. This often centres around Ray Kurzweil's ‘Singularity’, the idea that AI will at one point transcend human intelligence and become uncontrollable. Other AI experts like Stuart Russell, Brett Frischmann and Nick Bostrom have speculated at length on ways in which runaway AI could be a threat to our species. Although there are possible scenarios where runaway AI will lead to our demise as a species, this is not an issue that should worry us much in using AI for learning. Many, such as Stephen Pinker, Daniel Dennett and other serious researchers in AI are sceptical of these end-of-days theories. In any case, it is highly unlikely that AI for education will do much other than protect us from such scenarios.

Bias

Much more relevant is the topic of ‘bias’. The problem with many of the discussions around bias in AI, is that the discussions themselves are loaded with biases; confirmation bias, negativity bias, immediacy bias and so on. Remember that AI is ‘competence without comprehension’ competences that can be changed, whereas all humans have cognitive biases, which are difficult to change. AI is just maths, software and data. This is mathematical bias, for which there are definitions. It is easy to anthropomorphize these problems by seeing one form of bias as the same as the other. That aside, mathematical bias can be built into algorithms and data sets. What the science of statistics, and therefore AI, does, is quantify and try to eliminate such biases. This is, essentially, a design problem, and I don’t see much of a problem in the learning game, where datasets tend to be quite small, for example in adaptive learning. It gets to be a greater problem when using a model such as GPT-3 for learning, where the data set is massive. It can literally produce essay-like content at the click of a button. Nevertheless, I think that the ability of AI to be blind to gender, race, sexuality and social class may, in learning, make it less biased than humans. We need to be careful when I comes to making decisions that humans often make, but at the level of learning engagement, support there’s lots of low hanging fruit that need be of little ethical concern.

Race

The most valuable companies in the world are AI companies, in that their core strategic technology is AI. As to the common charge that AI is largely written by white coders, I can only respond by saying that the total number of white AI coders is massively outgunned by Chinese, Asian and Indian coders. The CEOs of Microsoft and Alphabet (Google) were both born and educated in India. And the CEOs of the three top Chinese tech companies are Chinese. Having spent some time in Silicon Valley last year, it is one of the most diverse working environment I’ve seen in terms of race. We can always do better but this should, in my view not be seen as a crippling ethical issue.

Gender

Gender is an altogether different issue and a much more intractable problem. There seems to be bias in the educational system among parents, teachers and others to steer girls away from STEM subjects and computer studies. But the idea that all algorithms are gender-biased is naïve. If such bias does arise one can work to eliminate the bias. Eliminating human gender bias is much more difficult.

Transparency

It is true that some AI is not wholly transparent, especially deep learning using neural networks. However, we shouldn’t throw out the baby with the bathwater… and the bath. We all use Google and academics use Google Scholar, because they are reliably useful. They are not transparent. The problems arise when AI is used to say, select or assess students. Here, we must ensure that we use systems that are fair. A lot of work is going into technology that interprets other AI software and reveals their inner workings.

Dehumanisation

A danger expressed by some educators is that AI may automate and therefore dehumanise the process of learning. This is often within discussions of robot teachers. I discuss the fallacy of robot teachers in the book. It is largely a silly idea, as silly as having a robot driver in a self-driving car. It is literally getting the wrong end of the stick, as AI in learning is largely about support for learners. Far from dehumanising learning it may empower learners.

Employment

The impact of AI on employment is a lively political and economic topic. Yet, before Covid, we had record levels of employment in the US, UK and China. There seems to be a fair amount of scaremongering at learning conferences, where you commonly see completely fictional quotes, such as ‘65% of children entering primary school today will be doing jobs that have yet to exist’. Even academic studies tend to be hyperbolic, such as the Frey and Osborne (2013) report from Oxford University that claimed ‘47% of jobs will be automated in the next two decades’. Seven years in and the evidence that this is true is slim. What is clear is that skills in creating and using AI for learning will be necessary. Indeed, Covid has accelerated this process. I categorise and list these new skills in the book.

Conclusion

I touch upon all of these issues in the book and stick to my original premise that AI is ‘not as good as you think it is and not as bad as you fear’. Sure there are ethical issues, but these are similar to general ethical issues in software and any area of human endeavour where technology is used. It is important not to see AI as separate from software and technology in general. That’s why I’m on the side of Pinker and Dennett in saying these are manageable problems. We can use technology to police technology. Indeed AI is used to stop sexist, racist and hate text and imagery from appearing online. Technology is always a balance between good and bad. We drive cars despite the fact that 1.3 million people die horrible deaths every year from crashes and many more have serious injuries. Let’s not demonise AI to such a degree that its benefits are not realised and , as I discuss in the book, in education and training the benefits are considerable.

 

AI for Learning

The book ‘AI for Learning’ is available on Amazon. In addition to ethics it covers many facets of AI for learning; teaching, learning, learning support, content creation, chatbots, learning analytics, sentiment analysis, and assessment.

 

Bibliography

Bryson, J.J., Diamantis, M.E. and Grant, T.D., 2017. Of, for, and by the people: the legal lacuna of synthetic persons. Artificial Intelligence and Law, 25(3), pp.273-291.

Kurzweil, R., 2005. The singularity is near: When humans transcend biology. Penguin.

Russell, S., 2019. Human compatible: Artificial intelligence and the problem of control. Penguin.

Clark, D., Review of Human Compatible https://donaldclarkplanb.blogspot.com/search?q=Human+Compatible+by+Stuart+Russell+-+go+to+guy+on+AI+-+a+must+read..

Frischmann, B. and Selinger, E., 2018. Re-engineering humanity. Cambridge University Press.

Clark, D., Review of Re-engineering humanity

https://donaldclarkplanb.blogspot.com/search?q=Frischmann

Bostrom, N., 2017. Superintelligence. Dunod.

Pinker, S., 2018. Enlightenment now: The case for reason, science, humanism, and progress. Penguin.

Dennett, D.C., 2017. From bacteria to Bach and back: The evolution of minds. WW Norton & Company.

Clark, D., Review of From bacteria to Bach and back

https://donaldclarkplanb.blogspot.com/search?q=Dennett+-+why+we+need+polymaths+in+the+AI+ethics+debate