Rowan is the co-founder and CEO of Pattern. He has been involved in a number of technology start-ups, and has also held senior executive roles in both the technology, and technology marketing sectors.
The ethics of technology
If we’ve learned anything in the last 18 months, it’s that technology, much like US politics, doesn’t police itself.The story that started the flood, was Cambridge Analytica, where journalists discovered that the UK based consultancy linked to Donald Trump’s presidential campaign had inappropriately accessed data from 87 million Facebook users. Over the course of 2018, a number of other data breaches were reported on, with perhaps the most concerning being Facebook’s data sharing arrangements with other corporate giants including Netflix, Amazon, and Yahoo.
Even more recently, He Jiankui’s announcement that twin girls had been born using CRISPR edited genomes has attracted intense public scrutiny. He’s leap into the unknown opens up many questions, not just related to his non adherence to scientific norms, but to the ethical dilemmas that may soon face would be parents, in a Gattaca-like dystopian future.
Although most of us don’t work quite so close to the cutting edge as He Jiankui, it certainly feels that the discussion of ethics in business technology has moved very much onto the front burner in boardrooms around the world. I believe that the onus is now on us to both reach an understanding of acceptable business practice, and then adopt the tools, checks and balances required to deliver on it.
At a minimum there are three pillars that need to be considered within most organisations.
At the centre of the Cambridge Analytica scandal is the sovereignty of our personal data, and it’s the storage, security, and tracking of this data that has legislators trying establish some level of control. However, it is businesses themselves that should be deciding independently of legislation - what is appropriate behaviour.
With data being bought and sold, on both open and closed markets, it seems clear that an individual’s right to privacy is at risk. Whilst GDPR and other initiatives go some way to mitigating this, individual privacy remains an attractive bauble for unethical businesses to pursue through technology.
There is a saying that, trust arrives into town on foot, and departs on horseback. With so much riding on trust, one rogue player can easily destabilise an entire industry. One of the most powerful ways to mitigate loss in trust, is transparency. As users we want to be confident that the data we have shared is being used in an ethical, and transparent way.
Our staff need to be well trained on ethics and accountability, with our engineers and product development teams understanding that these principles need to be baked into products and services. What we’ve seen recently, is that it’s much harder to retrofit ethics into products once they’re in-market. Like security, ethics and transparency need to be engineered in from the get go.
Another critical area of product development is the elimination of bias especially in the AI field. As product owners and developers, we need to be testing for bias and fairness as part of the development lifecycle. Perhaps the easiest way of achieving this, is to ensure that your team reflects the diversity of your user-base. In much the same way as we already test usability, accessibility, and security in advance of launching a product, tests for bias and fairness should now be included as standard.
As business already knows, self regulation is almost always preferable to legislation, and there's already a big movement toward ethical data practices that has been sponsored by the tech industry. It has become abundantly clear that legislators struggle to keep up with the pace of change in the industry, and therefore some kind of terms of reference framework is going to be required to assist boards, executives and developers understand how to create, build, and manage the products and services we take to market in future.
The horse has bolted, but that's probably okay. The barn was only version 1.0.