Move fast, but do not break too many things
When he founded Facebook, Mark Zuckerberg coined the motto "Move fast and break things" as a way to motivate creative design processes within the company. As the social media behemoth became a poster child for Silicon Valley’s negatively perceived attitude towards disruption, Zuckerberg changed the motto to (the far less catchy) "Move fast with stable infrastructure."
Nevertheless, it remains that social media and AI are extremely disruptive to society, politics and the economy, in both positive and negative ways. So it seems prudent to proceed more slowly, in order to allow for problems to manifest themselves at a speed that allows for their timely detection, diagnosis and repair.
However, there is substantial competitive pressure to move fast. If your company does not move fast in developing its AI technology, then your competitors might easily take over—and the collateral damage will happen anyway!
One solution to this problem is government regulation that applies to all companies developing AI. But often, regulations lag substantially behind technological innovation. So by the time suitable regulations are introduced, the damage is already done. Regulation introduces another problem: it can slow down innovation. So countries and regions that are heavy on regulations—e.g. the European Union—might become less competitive compared to countries with loose regulations.
Finally, given the moral imperative to progress AI technology for the benefit of humanity, slowing down also has a cost. Even if we consider only one AI application, namely driverless cars, it is plausible that delaying technological development and adoption could lead to hundreds of thousands of unnecessary deaths.
So how can we continue to move fast, while breaking as few things as possible? The answer is in increasing the speed of oversight, rather than slowing down innovation. Just as governments support the development of AI technologies, they can provide funding and legal backing for scientific studies that explore the potential impact of AI technologies on society, rather than leaving this type of research for companies to do on a voluntary basis.
References
Taneja, H. The era of ‘move fast and break things’ is over. Harv. Bus. Rev. 21, (2019).
Baer, D. Mark Zuckerberg explains why Facebook doesn’t ‘move fast and break things’ anymore. Business Insider 2, (2014).
Rahwan, I. et al. Machine behaviour. Nature 568, 477–486 (2019).
Kalra, N. & Groves, D. G. The Enemy of Good: Estimating the Cost of Waiting for Nearly Perfect Automated Vehicles. (Rand Corporation, 2017).
Aghion, P., Bergeaud, A. & Van Reenen, J. The Impact of Regulation on Innovation. https://www.nber.org/system/files/working_papers/w28381/w28381.pdf (2021) doi:10.3386/w28381.
Ledford, H. How Facebook, Twitter and other data troves are revolutionizing social science. Nature 582, 328–330 (2020).