Confronting Technological Change

When I was young, our family house was only a hundred metres or so away from the railway line. As well, my father worked for Queensland Rail for many years. Consequently it wasn’t surprising that I took an interest in trains, rolling stock and all that goes on above and around the rail tracks.

In my recollection my father had a couple of roles.

The first was as a “washout man”. In the days of steam trains it was necessary every now and then to clean the furnace of the steam engine. After allowing the furnace to cool, the washout man cleared it of ash and cinders and then washed it out and prepared it to be refired.

His second job was as an “examiner”. Despite the exalted title it was his job to go around the undercarriages of the wagons and carriages (called bogies) to inspect the wheels and the stuffing boxes (a type of bearing around the axles filled with greased wool). He had a little hammer which he used to tap the wheel tyres to ensure their integrity. The tyres were shrunk fit over the wheels and sometimes would work loose. He felt the stuffing boxes to ensure they weren’t overheating. (This could sometimes be dangerous because a failed stuffing box could sometimes render its casing red hot!)

There were many such jobs in the railway and petty demarcations were rife. My father was a traditional unionist and believed in the ethos of “one man, one job”. Therefore there was little ability to multiskill or share work around a team of employees. Consequently the work was unnecessarily labour intensive. What’s more Labor Governments ensured that the Railways provided generous concessions to railworkers and pensioners. I make no judgment on that, but as a result running the railways was unduly expensive and provided a huge impost on Queensland taxpayers which has only begun to be reduced in the last decade or so.

(The demarcations were so disruptive that when building rail carriages, as the rail workforce did in those days, a different category of workers was used to work on the insides of the carriages to the outsides. This resulted in the ridiculous situation that if an inside worker put a bolt through the wall of the carriage, it required an outside worker to do up the nut!)

But then there came a major disruption – diesel electric locomotives (commonly called diesels). This occurred from my recollection in the early 1960’s. Instead of a steam engine providing the motive power, diesels had a diesel engine which drove an electric generator which supplied an electric drive to power the train. This sounds unnecessarily complex but it had many technical advantages. It was also very disruptive.

Steam trains were fuelled by coal and the introduction of diesels caused some distress to the mines that provided the source of fuel.

The boilers that provided the steam to provide the motive power consumed a lot of water. Consequently there were many rail-sidings between the major stations that enabled the water supply to be replenished. These were subsequently shut down.

And in their usual way of QR and its petty demarcations, with the steam trains there were two people required at the front of the train – a fireman, who shovelled coal into the furnace, and a driver who worked the throttle to determine the speed of the train.

Of course diesels didn’t need a fireman nor also a washoutman.

This is a good example of disruptive technology. It changed the nature of the workforce. It led to faster trains with bigger payloads. It improved the economics of the rail business. But it caused some people to confront change that was uncomfortable to them.

The two factors that have contributed most to the improved standard of living we have enjoyed for many decades now have been the removal of trade barriers exposing our economy to international competition and the adoption of new technology that has enhanced productivity.

Of course there have been casualties and that is unfortunate. But if we fail to front up to free trade and new technology, we all in the end pay a price.

We must certainly work to manage the disruption that technology brings to protect those that are effected as much as possible, but not at the expense of eschewing the technology.

One of the prevailing myths that we need to overcome that emanated from the mid-twentieth century is the notion of “jobs”. With this myth came the belief that everyone was owed a job and once employed an employee owned his/her particular job and that the employee was somehow entitled to that job for life. Employees jealously guarded their job and often remarked they didn’t want anyone else doing “their job”. Often in order to guard their job they were reluctant to train others. And of course with this mentality, demarcations were ruthlessly applied.

(In my career as a power station manager I was to confront the worst of these demarcations. In unionised workplaces to do a job of any complexity multiple classifications were required. For example if a pump motor failed, to repair it required a mechanical tradesman to uncouple the electric motor from the pump, a rigger to place a sling around the equipment, and a crane driver to lift the motor out for subsequent repair. The logistics around such work was mindboggling. Not only did you need to have access to such people but they had to be scheduled to be there in the right order. As you might imagine, this generally resulted in people hanging around doing nothing because one of the critical classifications was temporarily unavailable. The productivity of such workplaces was appalling and the job satisfaction severely curtailed.)

It is more useful to view an enterprise, not as a place where many jobs are performed but a place where work needs to be done. To do this work efficiently the employer must of necessity employ sufficient workers with sufficient skills to do all the work that is required. And if the enterprise is to be competitive there should be no demarcations other than those based on skills requirement – i.e. any person can do whatever work for which they have the skills.

In my lifetime the nature of work has changed dramatically. The impact of computers, automation, robotics and artificial intelligence (AI) is ubiquitous. (Changes in communications technologies and social media have also made major impacts.) But the technological changes have had the greatest impact on labour intensive industries. Those hardest hit have been the unskilled members of the blue and white collar workforces. This has reduced the employment prospects of unskilled males more than any other segment of the workforce.

I have argued in other essays that the Industrial Revolution was a hugely disruptive change resulting in a large movement of employment from agriculture to the factories and mills. But fortunately the transition was made easier because the skills requirements of the new jobs created were such that they could easily be filled by agricultural workers.

The twentieth century move in Australia from manufacturing to the knowledge industries did not allow such an easy transformation because the skills requirements of the new jobs were vastly different from the jobs they displaced. Consequently the successful transition for an individual worker generally required reskilling. This has been a particular problem for older workers who are less confident of attaining new skills and it is they, more than others, who have resisted these changes.

Employee resistance to change is perhaps understandable but it can also prove very costly. In my career I saw many organisations spend huge amounts of money installing new computer software always justified on the productivity gains such software would allow. Yet those productivity gains were always premised on the fact that employees would be required to do things differently. Time and again I have seen such gains greatly diminished as employees found ways to do the work in the way they always had. Changing software for major systems is indeed extremely difficult, but changing work practices to benefit from such software change is far harder!

So whilst we have experienced major benefits from technological advancement in terms of improved standards of living, it is undeniable that there have also been casualties along the way. But hasn’t this always been the case? Consider the disruption of moving from horse-drawn to mechanised transport, or the introduction of mechanical harvesters into agriculture.

Then we must conclude that technological advances are inevitable, and that our welfare is enhanced by embracing them. But saying that, we are still obliged to do what we can to aid the successful transition of displaced workers.

Whilst I am promoting the necessity of Australia to embrace new technology, it doesn’t mean we have to be pioneers. It is often better to let others sort out the issues associated with new technology and then to be an early adopter.

Now I would like to discuss two particular, problematic technologies.

The first of these is nuclear technology. Australia, as I recall, is the world’s third biggest exporter of uranium (behind Kazakhstan and Canada), and we have vast quantities (about 33% of the world’s known reserves) of this nuclear fuel. Whilst much of the response to CO2 abatement has focussed on renewable energy technologies, the amount of renewable capacity is dwarfed by the burgeoning nuclear capacity, particularly in Europe and latterly in Asia.

According to the World Nuclear Association 11% of the world’s electricity is generated from nuclear power stations from about 450 nuclear reactors.

But, paradoxically, nuclear power generation is prohibited in Australia.

In the 1970’s an anti-nuclear movement developed that was initially concerned about nuclear weapons testing in Australia. It also sought to limit the mining and export of Uranium. This stance was vigorously supported by the ALP, Australian Democrats and the Greens. The Commonwealth Protection and Biodiversity Conservation Act 1999 in section 140A specifically prohibits nuclear power generation. Various state legislation also has the same intent.

No doubt the early nuclear disasters at Three Mile Island (1979) and Chernobyl (1986) provided huge incentives for caution to be taken in adopting nuclear technology. {The later event at Fukushima (2011) was less disastrous and was more instructive about proper site selection than anything else.}

But the proliferation of nuclear power stations, many of which have now run safely for decades suggests that we have little to fear now from this technology. Technical developments in recent times have continued to make such plant safer.

The proliferation of renewable technologies in Australia, aided by subsidies emanating from government –set renewable energy targets, with their intermittent outputs, has thrown greater reliance on our aging coal-fired fleet of generators to maintain security of supply. To be able to provide security of supply, the electricity network requires considerable despatchable capacity. This role can only be achieved by coal-fired, gas-fired, hydro and nuclear generation or by vast storage capacity using either batteries or pumped-storage hydro.

Unfortunately, if Australia is to abide by its obligations under the Paris Agreement (which is contentious but something I might save for another day) coal-fired and gas-fired generation adds to our carbon emissions.

There are very few viable hydro sites left in Australia and those that are viable are mainly remote from the grid.

There is little likelihood that battery storage in the capacity required to be effective will be economic in the foreseeable future.

Pumped-storage hydro is only viable when there is a large discrepancy between the maximum and minimum prices over the daily cycle in the electricity market. Pumped-storage facilities buy electricity from the market when the price is cheap, use that energy to pump water up into its storage, and then use that water to provide electricity generation when the prices are high. Because considerable energy is lost in the pumping phase there has to be a large price differential to make it profitable to then use that water to generate electricity. Without going into the technicalities, electricity demand has become a lot more constant over the daily load cycle which suggests it is likely that the price differential used to justify pumped-storage is likely to continue to reduce.

Consequently there is a strong argument that nuclear generation would be a beneficial addition to Australia’s generation mix. It provides base-load generation with no carbon emissions. We are well blessed with abundant fuel resources. It is time we took advantage of this technology which is now well-tested and demonstrable safe. This is one technology where emotive ideological stances have caused us to ignore the science and its broad, successful application by other nations.

Our history with respect to the adoption of nuclear technology indicates that there are often other considerations to consider in the adoption of technologies than just the science. My second example exemplifies the difficulty and complexity that can follow new technologies. The technology I have in mind is artificial intelligence (AI).

AI was initiated in the 1950’s. The first stage involved the development of neural networks which facilitated the creation of “thinking machines”. We can all remember the well-publicised contests between chess masters and computers. But machine learning really took off from the 1980’s. In the last decade it has accelerated with faster, more powerful computers and the advent of “big data”. (In my latter years as Chair of the Asset Institute we researched the many applications of big data to asset management,)

With these developments machines, are now beginning to make significant decisions beyond the routine operation of plant and equipment and data analysis.

The dilemmas that arise from this are now particularly highlighted by the advent of driverless vehicles.

Now this is not a new thing. Airplanes fly largely without substantial pilot intervention. Driverless trains are commonplace. On mine sites there are many driverless trucks.

Airplanes have imbedded intelligence to take avoidance action if other aeroplanes approach too close for comfort. But although airplane collisions produce disastrous outcomes, in most circumstances there is little likelihood of that happening. But driverless cars face road conditions where accidents are far more likely to happen. In the face of a likely accident a vehicle can take a range of options to avoid damage or to minimise human casualties. In such a situation moral dilemmas appear. Consequently the decision making of the machine needs to be programmed to ensure the appropriate decision is taken.

(I hasten to add that this is not a problem exclusive to driverless vehicles. As outlined in a recent article by software manufacturer SAS the problem is far more extensive. In arguing why safety of AI systems is paramount, they wrote:

Whereas it may be little more than minor nuisance if your laptop crashes or gets hacked, it becomes all the more important that an AI system does what you want it to do if it controls your car, your airplane, your pacemaker, your automated trading system or your power grid.)

This problem is often misunderstood. Driverless cars are often called autonomous vehicles, but they are not autonomous in the conventional sense because the decision making protocols are determined for them by human agents. The intended outcome is to ensure that the goals of the machines are aligned with ours. But the dilemma is whether those goals should be determined in favour of the individual to whom the machine is responsible or whether the goals should be more altruistic.

Returning to the driverless vehicle, making such a decision may have starkly different outcomes. When driving a car and an accident is imminent most drivers will act out of self-defence and take action to minimise the likelihood of self-harm.  But with driverless vehicles, the protocols are likely to try to ensure that the minimum damage ensures. This might sometimes put the passengers at risk to ensure a bigger calamity is avoided.

Some have suggested that that the progress towards driverless vehicles should be halted until there is agreement on the morality of the decision making algorithms.

At the recent NSW Labor conference the ALP considered regulating the introduction of AI because of its concern about resulting job losses. They expressed concern at a recent study by the Committee for Economic Development of Australia which suggested that in the next ten to fifteen years 40% of Australian jobs might be impacted by AI. The conference also endorsed a motion to ensure an ethical framework was in place before allowing the widespread use of driverless vehicles. The German government has developed ethical guidelines for decision making by driverless vehicles and the Singapore government is also developing such a code.

It was no surprise that this approach was led by the National Secretary of the Transport Workers Union, Tony Sheldon. Most of the members of Mr Sheldon’s union are truck drivers. Widespread adoption of driverless vehicles would decimate his membership so it is inevitable that Sheldon and his union would oppose such a development.

Nevertheless it is prudent that an ethical code be developed to guide the decision making of driverless vehicles. This is one technological change we should approach with caution. But we should not be afraid to embrace it when the ethics of the decision making are sorted. After all machine decision making is more reliable than human decision making. If we can sort out the protocols we should not fear driverless vehicles. Humans are far more fallible than machines.

So then let me summarise my position with respect to technological change.

It is my contention that in general we should embrace technological change because overall it enhances productivity and leads to a higher standard of living. Eschewing technological change might sometimes postpone uncomfortable structural adjustments but it is largely postponing the inevitable. Putting off technological change often makes the transition, when it inevitably arises, even harder.

Nevertheless such change will often bring casualties and we need to be diligent in easing the pain of change for those so affected. The burden of change often falls disproportionately on older members of the workforce who often doubt their capacity to learn the new skills required of them.

We need to be rational and not emotive in responding to technological change. As I have outlined above our refusal to embrace nuclear technology is not rational. It is now a proven, safe technology which, if embraced, would undoubtedly benefit our economy.

But as I have outlined in my discussion about AI, we should also be cautious in embracing technologies that might have unintended deleterious consequences.  As I have argued above, it is often prudent not to be a first mover but to be an early adopter once the consequences of implementing a new technology are well established.