I am one of the ethnographers of the NGI Internet of Humans community. But, in my spare time, I occasionally write about technology. In particular, I am very interested in AI governance and its use by governments.
I recently co-published a piece in the Dutch Volkskrant about AI use by the Dutch government, you can find it here in Dutch:
and here in English:
Would love to hear what people think, but also how your government is using AI/ML systems and what the consequences are as well as what justifications drive the integration of AI into government provision of services and goods?
Excellent question. I am a double citizen, but in this post I will consider “my country” to mean Italy, the country where I was born.
Italy has created an AI Task Force as part of the Agenzia per l’Italia Digitale. It consists of academics, practitioners and a few free thinkers, rather than civil servants (I know and respect several of them), and seems to be tasked mostly with providing a vision and principles. In March 2018, the Task Force released a white paper which makes all the right noises (human-centric! bias-free! inequalities reducing!), but it’s hard to know how credible these noises are. In the past, the Italian state has not always been a good purchaser of ICT services. Vendors have had a lot of torque in pushing through proprietary solutions; and even when the government-side buyers had a clear idea of what they wanted, they have tended to be motivated by cost and staff reduction. For example, as in other countries, staffed counters to serve citizens trying to access public services are disappearing everywhere, replaced by websites and phone apps. This makes sense (especially if you do not to question the administrative ordering underneath, but this is another story that James Scott and David Graeber have already told very well), but it raises major inequalities problems. Senior civil servants know this, but they are under pressure to reduce their workforce and save costs.
The Task Force is also trying to inventory government projects that use AI, as well as simple needs that could be met by deploying AI tools. The inventory (rather underwhelming) is here.
California still dominates in word and in thought and encourages the concept of a single way, technological deterministic approach. If the development of artificial intelligence is fully shaped by private stakeholders, based abroad France and Europe will have no other choice than to their vision. This is already happening in the public sector. Think of the agreement signed between Microsoft and the Ministry of Education during the previous five-year term and the DGSI’suse of software provided by Palantir—a startup with links to the CIA. This is equally true in the private sector. Across Europe, businesses convinced that they have already lost the battle frequently succumb to the persuasive powers of the U. S tech giants, sometimes at the expense of our own digital “nuggets”. […] Considering that France and Europe can already be regarded as “cybercolonies” in many aspects, it is essential that they resist all forms of determinism by proposing a coordinated response at European level.
Conclusion:
This is why the role of the State must be reaffirmed: market forces alone are proving an inadequate guarantee of true political independence. In addition, the rules governing international exchanges and the opening up of internal markets do not always serve the economic interests of European states, who too frequently apply them in one direction only. Now more than ever, we have to provide a meaning to the AI revolution. (emphasis mine)
France has the historical role of pushing Europe to do things its own way, rather than passively follow America (or, in the present day, China) around. If the French get behind this position, we have a real opportunity for Mariana Mazzucato’s vision of mission-oriented innovation to become Europe’s battlecry in AI (and, why not, in Internet technology in general).
I only read the executive summary of the report, but it seems sharper than the Italian one, and it contains some actionable ideas. But its main value is, of course, in signaling political commitment. I came across it after one of my favorite scientists, Cesar Hidalgo (director of MIT’s Collective Learning group), announced on Twitter he’s moving to Europe, citing the Villani report as having inspired the creation of a new Institute at U Manchester:
“If you are coming from silicon valley, don’t bother applying - we already know you have no ethics”- overheard in a conversation where someone was recruiting an ethical advisor for their tech company.
