Select a result to preview

Andrews2019PublicAdministration

Leighton Andrews, "Public administration, public leadership and the construction of public value in the age of the algorithm and ‘big data’"

Bibliographic info

⇒ Andrews, L. (2018). Public administration, public leadership and the construction of public value in the age of the algorithm and ‘big data’. Public Administration, 97(2), 296–310. https://doi.org/10.1111/padm.12534

Commentary

⇒ In general, what about this text makes it particularly interesting or thought-provoking? What do you see as weaknesses?

What I find interesting is that AI and big data are in this article discussed as "wicked problem", which is an important concept in Public Administration. A wicked problem is a problem characterized by high uncertainty of information, as well as a lack of shared values on the problem, therefore it's extremely difficult to solve. In a sense, this is the PA equivalent to an ethical dilemma. In this article, the author therefore treats dilemmas and problems inherent in the use of algorithms in governance by likening it to wicked problems. This is especially true because public value theory (PVT) is used in this article as argument for ethical / moral underpinning of arguments in discussions about algorithms and AI. Usually in PA, wicked problems call for more research and frame alignment of issue and value perceptions, in order that policy may be formulated to deal with the problem. From this context, I think it's a weakness the author does not present a hard call for governance on AI, even stating that some algorithms are covered with current legislation. If the author agrees there is lack of information and lack of shared values, with the possibility of AI to develop very quickly, it will not work to react in hindsight. PVT is a governance process, if this is used in discussions about algorithms, then this notion of governance must be present in solving AI problems as well. I therefore think the author has conflicting ideas and statements in this article, and thereby negates some of his own points.

Excerpts & Key Quotes

⇒ For 3-5 key passages, include the following: a descriptive heading, the page number, the verbatim quote, and your brief commentary on this

Quotation here: "Pasquale Pasquale2015BlackBox says ‘authority is increasingly expressed algorithmically’. Yeung (2017) speaks of ‘algorithmic power’. This is not, of course, to say that all algorithms require governance or regulatory intervention. Gillespie (2014) has called certain kinds of algorithms ‘public relevance algorithms’ which have ‘political valence’. "

Comment:

I chose this quote to highlight the inconsistencies in the article, because when reading this it immediately struck me as an internally conflicting statement. In the article, the author explains there is always contestation. This is especially true of wicked problems where there is contestation both of values and information. In Public Administration it's accepted that wicked problems always require governance, therefore I think it's very confusing and conflicting to not expect the same for AI. Especially with an eye to the future, to be governance ready for AI means not falling back on older legislation, but trying to stay ahead as well. All AI that can make decisions and influence lives has public relevance, as everyone is increasingly confronted with it.

Quotation here: "regulation is never neutral: as Moe (1990) said, ‘for most issues, most of the time, a set of organized interest groups already occupies and structures the upper reaches of political decision making’. He suggests that compromise is often built into the construction of regulatory institutions, whether they are agencies or laws. Governance and regulation develop in a contested context."

Comment:

This quote is especially convincing for my criticisms, because if in PA we agree that regulation, governance, and policy are never neutral, why is this not accepted for AI? Algorithms take on the biases of the creator, therefore especially in the public domain this can have tremendous implications, think f.i. about systemic racism that is already (maybe even unknowingly) built into algorithmic decision making. Therefore, governance of algorithms is absolutely necessary.

Quotation here: "The overall conclusion from the Royal Society/IPSOSMori research and further survey evidence from the EU’s Digital Single Market programme (European Commission 2017) suggests that people are open to exploring the role of artificial intelligence, although they believe that these technologies require ‘careful management’."

Comment:

I believe this quote illustrates a societal call for governance of algorithms and AI, in order to legitimize and gain more trust in algorithmic governance. With a democratic system we trust the government to act in our best interest, but clearly we don't automatically trust the same for AI, rather we are more skeptical. Therefore governance of AI is the only way to legitimize AI outcomes and processes in a democratic system.
#algorithmicGovernance, #democracy