To Support Innovation We Need To Accept Government Failure

The most insightful idea that I have heard on how best to support and encourage new companies to create innovative products is this: companies don’t need a grant or advice. They need a customer. Therefore, the best thing that government can do to encourage innovation is buy products from emerging companies. Unfortunately, the incentives of government encourage the exact opposite, and so we need to change this.

Incentives of Government

There is no reward for a bureaucrat who takes risks and achieves innovative things. You cannot get fast-tracked from customer service to the head of a department by inventing a new way of organising the department’s files. Unlike Google, where Larry Page and Sergey Brin became some of the richest and most influential men in the world for organising files on the internet. Instead, the strategy for success in the public service is an avoidance of any long-tail risks, i.e., mistakes. Or, more accurately, any left side (downside) of the distribution risks at all! The risks to avoid are not the catastrophic failure one can experience in the private sector (e.g., bankruptcy) but even smaller and comparatively inconsequential errors. A repeated breaking of rules or systems is usually a public service career-limiting move. And the reason is that ultimately these rules against minor downside durations are to prevent public embarrassment of the Minister responsible for the department.

If a department makes a mistake such that it makes the news that can cause embarrassment for the responsible Minister and this puts the Minister at risk of losing their job (through future election loss, or anticipated loss), and so the Minister will require the department to prioritise “not making any perceived mistakes” above all else, including competency.

A frustrating example is the failure of Governments to send communications through modern means, especially text messages or emails. Because of the perceived risk of personal information going astray on the internet, 95% of physical letters today are sent by Government (and also businesses that operate with the modernity of Government). Apparently, “emails are too insecure.” Well, yes, someone could hack into my emails, but they could also open my letter box and take its contents. For important matters, I ensure that they are sent to my secure post office box, and so too, should I be able to choose to receive material at a more secure email address. Or perhaps someone doesn’t receive correspondence that they are especially guarded about. They should be able to choose to receive it via their general email account protected by “password1234” just as they have their letters spill out of an unlocked post box.

But what matters is not extant security, merely the perception of security. I recall trying to get some urgent (and private) client material out of the ATO, who refused absolutely to send it to my firm email address, accessible only by me, but were happy to fax it to a fax machine located in a waiting room where members of the public could freely enter and access. I had to stand next to that fax machine for an age and wait while it slowly received and printed documents – then I had to scan in and email to the client in any event.

I think even worse is when some department decides to build its own “secure” online communication portal. Firstly, because they are not skilled, motivated, or incentivised to create a form of communication useful to me in the way that a private company that must convince people to positively use their software must. Secondly, because the last thing that I want is to have to monitor yet another communication platform, and one that is going to be clunky to operate, difficult to access, and is mostly going to be fitted with messages reading “Advance notice of system outage for maintenance.”

Accepting Failure

It is environments such as this that have given rise to the adage, “no-one loses their job by buying IBM.” That is, purchasing from an established and reliable (but by implication less innovative) vendor for a product may not lead to the use of the most advanced technology being implemented, but it will not lead to the purchase of a problematic product that could ensure negative publicity. (Quare whether buying from tech giants is actually less innovative, given that they – including IBM – are the source of the largest quantity of AI infrastructure and research).

This sentiment of risk aversion leads to a slow adoption of new technology – even from established tech companies – and almost entirely shuts out emerging companies. This is to the detriment of the delivery of government services.

For example, I know of a couple of departments that still routinely fax information between them, with information having to be typed in the first department’s computers, printed then faxed, received, printed, and then re-typed at the second department. So, we are not exactly talking about whether AI should be given decision-making powers. We are just trying to move to 1990’s tech.

In my view, we can both speed up the adoption of existing technology (and move away from fax machines) as well as adopt newer technology by allowing and even encouraging (limited) failure in Government. It is, of course, not helpful to have, say, everyone’s welfare payments fail to arrive or national secrets leak onto the internet. This is not about critical infrastructure. We should allow government to experiment with a limited proportion of their technology spend, say, 10%. They should be expressly allowed to purchase newer products knowing that a certain proportion of them might have some kind of risk.

An easy start could be allowing emails if the client opts into it. Sure, there is a risk of someone mistyping an email address and private information going to an unintended address. However, this small risk is well outweighed by the efficiency and convenience of receiving emails (and being at least in the 1990s). If the Minister sets out beforehand that there may be some small risk but that it is worth it because of the broader benefits, they can inoculate themselves from any potential mistakes.

Besides catching up on existing (old) technology, there is a plethora of opportunities for government to adopt and try more recent technologies.

Instead of trying to support emerging companies with grants, subsidies, advice, and hackathons, why not give them something that they need more: a customer. A customer is real cash flow (as opposed to a temporary grant) and can lead to improved business stability and validation for other customers. The department benefits through more efficient and useful technology. The public benefits from better and more convenient service delivery. All the Minister needs to do is jawbone the public that the risk of innovation is (limited) technology transgressions.

Faster, better, cheaper: embrace failure.