🔬 Research summary by Connor Wright, our Partnerships Manager.
[Original paper by Gianluigi Viscusi, Anca Rusu, and Marie-Valentine Florin]
Overview: Different nations are now catching on to the need of national AI strategies for the good of their futures. However, what really drives AI and whether this is in line with the current fundamental values at the heart of different nations is a different question.
Introduction
With the release of the UK Government’s AI strategy, the question of what is driving its design comes to mind. AI has long been touted as a sure way to improve public service delivery and administration efforts. However, questioning what kind of values are currently driving government initiatives has not been too visited. Are the private and public spheres motivated by different values? Does humanity itself enter into the AI conversation? Is AI being treated in too much of an instrumental maner? I will now explore these three questions in turn.
Key Insights
Do the private and public sectors differ?
With different end goals and different audiences, the private and public sectors can differ on many aspects of governance. However, “accountability, expertise, reliability, efficiency, and effectiveness” (p.g. 2) were found to be held in common between the two spheres. Other aspects like “professionalism”, “efficiency”, “openness” and “inclusion” have been common to both as well (p.g. 2). However, I believe what can differentiate the two are the interpretations of the values listed. Efficiency will differ from business to business, especially in terms of what is deemed as the threshold. Moreover, “inclusion” (p.g. 2) and who is subjected to it can vary widely in terms of extent, depth and what inclusion entails. Being included in the AI governance could range from participating in its design or just occasionally being informed on the changes being made to an AI. Most of the time, a lack of inclusion is witnessed.
A lack of focus on humanity
Principles such as “transparency”, “privacy” and “responsibility” are mentioned in AI governance strategies more so than “Human dignity” (pg. 4). AI is often touted in the media as the golden ticket towards greater ‘prosperity’ for humanity, but humanity’s role in this prosperity is often left untouched. To illustrate, there were fewer risks and challenges identified in the AI strategies of each country than there were values. So, what are these values contributing towards if governments do not articulate the problem? From my reading of the report, AI is far more accepted as a tool than as a socially designed and sensitive technology.
Instrumental normativity over social normativity
Phrases in the shape of privacy, efficiency and transparency often are preferred by governments instead of words such as democracy. One way to express this substitution is by acknowledging the tension between improving administrative features and simultaneously focusing on societal issues. Resultantly, the values held at the core of constitutions and manifestos are often sidelined when thinking about AI, like with the lack of reference to democracy. Questions then arise of whether values serve as sound guiding principles for AI at all, or are they another example of a tokenistic gesture.
Between the lines
What is crucial for me to consider is how what is valued varies so differently across different nations. As explained in one of our event summaries on AI in different national contexts, the cultural interpretation of different values can vary widely. However, fragmentation is not always a bad thing if each country adheres appropriately to their given interpretation.
The next topic of debate comes from how the fundamental values that the entities in the report expounded were not concurrent with the values shown in the AI strategies. Committing to AI values beyond simply writing them down is a known struggle within the space. For me, writing down the value and what problem this value, if championed correctly, helps to prevent is one way of contextualising and focussing efforts on the issue at hand. The more context and grounding in what AI is being deployed to do, the more the equally valuable social values we hold can appear.