AI and human values: but which humans?

Can artificial intelligence develop without bias, clearly evolving without a tendency to favour any one set of values over another? An Op-Ed piece in the New York Times makes its view clear enough in its choice of title: “Artificial Intelligence’s White Guy Problem”. From face recognition software that mistakes black people for gorillas through predictive software that operates on assumptions about recidivism that would provoke outrage if articulated by human agents: these egregious mistakes have at least done us all a service in alerting us to the fact that algorithms are not by nature free of moral considerations. They work as distinct outputs of the people who craft them, reflecting the choices and betraying the values of their creators.

Bias, however unintentional, proceeds from privileged perspectives that go beyond the assumptions of colour and race. The Times Op-Ed instances the gender-driven discrimination revealed in last year’s academic study of Google job advertisements, revealing that men were more likely than women to be shown adverts for the more highly-paid jobs. Elsewhere, there has been the instance of political bias that arose in the story reported in the Financial Times, concerning Facebook’s alleged search bias in favouring liberal over conservative views. That led to a story in Salon that asked if “Google results could change an election.”

Work is being done on what a thoughtful blog on the London School of Economics website refers to as “algorithmic fairness – a systematic way to formalise hidden assumptions about fairness and justice so that we can evaluate an algorithm for how well it complies with these assumptions.” But vigilant circumspection must ensure that humanity never leaves it to algorithms to protect human values. In spinning the wheels of privilege, it may seem to the (mostly) wealthy, white, older, male population that is driving AI’s development that humanity’s existential risks lie some way off in the future. The reality for the lesser privileged is that for them, the impacts of implicit AI bias are being felt now.

Brexit and brains: blink and miss the logic

Britain’s decision to remain in, or exit from the European Union has surprised us all with the extent of the passions that have been stirred up. In the context of Brains and Minds and the exercise of applied intelligence, what is interesting is the extent to which the dialogue between “Remain” and “Brexit”, as exemplified in last night’s BBC extravaganza of a “debate”, betrayed our species’ trademark rush to judgement in the blink of an eye, the exercise of a wide range of cognitive biases in defense of unabashed passion, and disrespectful trashing of the motives of those who disagree.

Neither in last night’s debate nor in the dialogue that has dominated the UK’s attention for weeks now, has there been any shortage of good questions and answers on both sides. Neither has had a monopoly on reason or passion. Nor indeed have we been short on “mongering” either. One charges the other with spreading hate on immigration; the reverse monger is of Project Fear on the economy. The debate’s closing statements featured a defence of expertise from Remain, followed by a visceral plea from a politician who never recognised a rabble he could not rouse, deploying an imperfect analogy in defense of a Brexit case that was holed before the debate began by one of its advocates actually dismissing the claims of experts.

What does it mean that a former education secretary with a reputation for the deft application of intellect should scorn people with expertise? With all the clamouring for good evidence and information about a question as big as the challenge of Brexit, and given our slow ascent from the swamp to the summits of human achievement, do we not want to take note of the people who have made a living from considering evidence and thinking things through on some relevant topic, earning their stripes and tee shirts along the way?

People whose minds will not have been made up can feel what they like in responding to populist spasms, but cannot fail to recognise in Remain the greater benefits of sound circumspection. Appeals merely to hope and faith and “getting our country back” are modes of thinking that are declining in the face of reason, science, and a belief in the need to work together for a better world.

Bear necessities of augmented intelligence

A favourite joke involves two barefoot guys in the forest who spot a bear off in the distance, but lumbering towards them and closing fast. They have to run for it but one stops first to put on his running shoes. The other laughs and asks if the running shoes are going to enable him to outrun the bear. Finishing his lacing up, the first guys smiles and says: “I don’t need to outrun the bear; I just need to outrun you.”

As to the anticipated challenges of matching the powers of Artificial Intelligence when it supersedes the capacities of what humans can do – when it graduates to “Super” status and leaves humanity floundering in its dust: we may take a lesson from the bear joke and wonder if there is a middle way.

It appears from all the media commentary that we have little idea now as to how to design SAI that will not obliterate us, whether by accident, or through its own malign design, or by some combination possibly exacerbated by the purposeful interventions of thoroughly malign humans. Can we at least get smart enough to do what we cannot yet do, and find a way of programming SAI to help us solve this?

Without getting into the philosophical difficulties of bootstrapping something like intelligence, two things are clear. We must get smarter than we are if we are to solve this particular problem; and brains take a long time to evolve on their own. We need an accelerant strategy, and it will take more than brain-training. Research must proceed more quickly in brain-computer interfaces, nano-biotechnology and neuropharmacology, and in the sciences of gene editing and  deep-brain stimulation. While research into several of these technologies has been driven by cognitive impairments such as movement disorders and the treatment of depression, their capabilities in areas of potential enhancement of cognitive function are attracting greater interest from the scientific and investment communities. It is definitely becoming a bear market.

Our candle flickers in the cosmic night

On the evidence of our own species, it seems that intelligent life cannot develop without evolving a capability for destroying itself, or at least for acquiescing in a decline through obsolescence into oblivion. What is additionally remarkable is how rapid the evolution from dust particle to dust particle can be, notwithstanding the brief sparkle of celestial fire that lights the passage in between.

This shelf life of intelligence needs to be borne in mind when running the Drake Equation on the likelihood of extra-terrestrial intelligent civilisations. It is not enough that they defy the odds to kindle themselves into existence. Their chance of connecting with any similar intelligence elsewhere in the universe will be slight if they cannot manage to stick around for somewhat longer than the mere slight smear of time in which Homo sapiens has illuminated its small corner of one little galaxy.

The huge number of intelligent civilisations that might be calculated on the Drake formulation become somewhat more meagre if they appear and disappear on the timeline that our own civilisation seems determined to follow. A truly universal telescope programmed to detect intelligent life might, over the course of all of time, pick up the flickering in and out of existence of so many thousands of smart civilisations like some vast constellation of fireflies, flaring up in their nanoseconds of existence and flashing amidst the vasty depths of the cosmic wilderness. If all civilisations play out like ours – at the time of writing occupying its perch atop Earthly creation for less than .0001% of our planet’s existence – then the chances of any two co-existing, and co-existing in something like the same galactic neighbourhood, must be small.

And when one ET does reach out successfully to us, what are the odds that its message – once filtered through the interstellar Rosetta Stone we have yet to invent – might say anything more meaningful than “Come quickly!”

Time has ticked a heaven round the stars

In the green fuse of the young Dylan Thomas’ imagination we find the perfect description of how humanity’s knowledge of the universe has proceeded from mute awe to a better but still imperfectly informed wonder. All it took was time – and science – and our notion of heaven was transformed from the clumsy metaphor of celestial theme park to something far richer, more vast and various, and beautiful beyond comprehension. And most wondrous of all, we humans are not only actively immersed within this heaven, albeit on the nanoscale, but we are conscious of being so, and of being so in the here and now.

It is easy for us to see today how religions ignite. While all other species seemed happy to proceed from meal to meal without any need for meaning along the way, Homo sapiens has sought explanations, patterns, and a sense of its place above and beyond the brutish rants and ruttings of everyone’s daily lives. Given what we knew about what we flattered ourselves to suppose was the universe two thousand years ago, it is not surprising that the revelations and rules comprising the Pentateuch, Bible and Koran emerged as the defining Operating Manual for Life on Earth. And in understanding more now about what we didn’t know then, and given our inbred venalities and credulity, it is even less surprising that these religions caught on.

With what has happened over the last two millennia – and in science and technology, what has transpired particularly over the last two centuries – it would be stranger if anyone were now to propose one of the “great faiths” as a credible belief system for today’s world. (Although, as a reminder of the limpet-like tenacity of human credulity, it is less than two centuries since the appearance of the “Book” of Mormon.) But on balance, our cognitive horizons continue to expand and, with them, our aspirations for new frontiers of intelligence and wonder in an enlarging universe. We may not wonder at the answers conceived by religion in the infancy of our species but, in our progress beyond I Corinthians 13:11 to the irony of John 8:32, we open up new vistas of potential in the flowering of human intelligence. The truth can indeed set us free, although perhaps not in the manner that the Jesus of scripture intended.

Control is certainly a human problem

One of the highlights of the CODE Conference that is concluding today in California has been the concern expressed by Bill Gates that, although these are exciting times for innovation and technological accomplishment, there are at least two considerable challenges. One concerns the threat that evolving AI represents to a wide spectrum of human jobs. The second is summed up in the sub-heading to the article linked to above: “The real challenge is ensuring humans stay in control.”

Is this really so? Maybe humans should retain control, but it is more than just a question for parlour, pub, or philosophy colloquium. A reflexive “Yes, of course they should,” is all too easy, but a slower and reflective “Yes, although . . .” would give us a chance to consider the assumptions we live with, and the direction that Life on Earth is taking, possibly independently of whatever it is we think we want.

How are we doing with the control we’ve got? In the miniscule percentage of 1% of the history of Planet Earth during which time humanity has been the governor, how have we done? And who are “we”? By humans do we mean a thoughtful, informed, judicious and circumspect collective of incorruptible and accountable grown-ups? Or are we talking about the United Nations? Or a chaotic inferno of babbling rabbles? If, alternatively, we are talking about smaller, indeed singular subsets of humanity, are we happy with the definitions of power that history has distinguished among the self-entitled 1%, or our kings, priests, plutocrats, and other monomaniacal psychopaths?

Must the limit of our ambition for Life on Earth be constrained by the limitations of our own humanity, or do we ourselves adapt in pursuit of a higher intelligence? Is it better that we retain control and fail, or can we give full reign to an emerging intelligence that assumes more executive control as we evolve a higher wisdom? It may be that “control”, as the watchword for the 100 millennia during which Homo sapiens has struggled for planetary dominance, is exchanged for “co-operation” as the key to a happier future for humanity and for the planet. And to the extent that control retains a role, it might best be applied to our own self-destructive impulses.