The spectre of inequality

Consultation as one of the ways to address Yuval Hariri’s latest warnings.

Consultation professionals are essentially concerned with process. We are not supposed to be pursuing our own political, social or economic agendas. Our role is to facilitate informed debate, though, over the years, many have argued that improved education, liberal values and the wisdom of crowds have, all other things being equal, produced more progressive outcomes. Without question, one of these involves equality.

By any historical comparison, many kinds of inequality have been addressed. Progress has been made to reduce, if not eliminate discrimination on the grounds of class, race, gender or even age. It has been a long and painstaking process, with so much still to be done. The UK is not exactly a laggard, but neither is it top-of-the-class. As we entered the 21st century, there was room for optimism that this egalitarian process would continue, or maybe accelerate. Although he is not the first, Yuval Noah Harari (who wrote the best-seller Sapiens a few years ago) has now argued that the reverse may happen!

Reading his book, 21 lessons for the 21st century, it is easy to become concerned. He focuses on Artificial Intelligence (AI) and on Bio-engineering and foresees that their combined effects will be to frustrate the moves towards greater equality. Decision-making by an algorithm is already proving a nightmare in the recruitment industry (by applying rules to analysing candidates’ CVs) and in banking (by ruling upon people’s fitness to be lent money), but we are but at the start of the AI age. To take another example, the emerging use of biometric sensors in our bodies tells the data-owner far more about incipient medical conditions than the patient knows himself/herself. Those with access or funding for this new technology will potentially enjoy immense advantages in securing health care.

There are consequences for freedom as well as for equality. When so much is decided by an algorithm, it becomes important to know who owns that formula and what accountability there might be for the biases and weightings it reflects. Similarly, the ownership of big data, whether about our bodies or the millions of driverless cars that will travel our future roads, also becomes crucial. If Google, Facebook or other private companies own these critical tools, that suggests one future scenario. If Governments own them, is there more accountability? In short, what kind of Big Brother do you most dislike?

Consultation has a role to play. There are ways to avoid non-human decision-making exacerbating social and economic inequalities. An algorithm can do good as well as cause harm, but that pre-supposes that whoever has the choice is open to influence and competing arguments. Having an open debate as to whether certain factors should or should not disadvantage particular groups or individuals is an important safeguard against the tyranny of the technology.

One of the most dangerous prospects is technology-adoption by stealth. Whether it is the replacement of call centre staff talking to claimants for Universal Credit by a computer than does it all automatically, the waiting list for routine hospital operations that suddenly becomes system-determined, or the new points-based immigration system – if it ever comes – the decisive moment is when it is designed, and approved.

Where significant changes are proposed to the NHS, there must be public and patient involvement. When rural fields on the outskirts of town are to be built upon, there must be local authority planning permission. Where new powers are given to a Government department, there has to be legislation of some sort in Parliament. But when human decision-making changes to machine code, are there equivalent safeguards? What Hariri and others point out is that even subtle changes to the ways in which public services are administered can, cumulatively change the balance of power between decision-makers and the general public. Do we, perhaps need to define a special trigger-point at which the involvement both of those that are impacted, and the wider public becomes necessary? Cue consultation!

Longer-term forecasts are indeed scary. Hariri says that “… authority could shift completely from humans to algorithms. Once AI is smarter even than the human elite, all humanity could become redundant.” More realistically, he observes that those who control the systems, the applications and the data become so much more powerful than the masses and “ …once the masses lose their economic importance and political power, the state loses at least some of the incentive to invest in their health, education and welfare. It’s very dangerous to be redundant. Your future depends on the goodwill of a small elite.”

To mark the start of the new decade, the last Tuesday Topic (No 355 A decade of debate) considered The Economist’s analysis of forthcoming technology and concluded that we would see a plethora of consultations as society collectively figures out the best policies to address some pretty big challenges. By focusing on just one aspect – a threat to equality by surreptitious adoption of AI and related advances - we now highlight the need for greater consultation on their detailed implementation. Our legislators need to be alive to the dangers and willing to mandate compulsory engagement with wider communities and groups if we are to avoid some highly unpleasant consequences.

TRIGGER POINTS

1. Have you yet begun to understand how new technologies will affect you, your work and your interests?
2. Should you read Hariri’s works. This is a link to the book that stimulated this Topic.
3. Here is a useful summary of the book.

The Institute has many training courses andservices that help organisations decide upon the kind of consultations as are appropriate.

This is the 356th Tuesday Topic; a full list of subjects covered is available for Institute members and is a valuable resource covering so many aspects of consultation and engagement.