Governments are increasingly approaching artificial intelligence with a religious zeal. In 2018, at least 22 countries from around the globe, as well as the EU, had launched national strategies to make AI a part of their businesses, while others had announced ethical frameworks on how AI should develop. The EU documents over 290 AI policy initiatives in individual EU member states between 2016 to 2020.
Ireland is the latest to announce its national AI strategy ” AI- Here for Good“. The aim is to be “an international leader when it comes to using AI for our economy and society through a people-centred and ethical approach to its adoption, development and use”.
Eight policy commands will be used to achieve this, such as increasing the trust and understanding in AI through the use of an “AI Ambassador” (a veritable AI High Priest) to spread the message throughout the country. A second aspect is to encourage AI adoption among Irish businesses and government in a moral and ethical framework.
The strategy praises existing AI-based applications that, for example, improve cycling infrastructures in Dublin, provide Irish-language tools, save energy, and comfort dementia patients. But it’s hard to see how these apps could double the economic growth.
AI is a key component of a select few digital platforms companies such as Google, Apple, Facebook, Amazon, and Alibaba, or GAFAA. Due to the large amount of data required by current AI, they enjoy winners-take-most advantages. The data becomes more profitable as your platform grows.
Digital platform firms disrupt existing businesses by a href=”https://www.amazon.com/Platform-Revolution-Networked-Markets Transforming/dp/0393249131″> outcompeting/a> them every time. A good example is how Google, essentially an online search engine, disrupted the advertisement-driven business model a href=”https://www.bloomberg.com/opinion/articles/201902-12/go Digital platform companies disrupt existing businesses every time by outcompeting. A good example is how Google, a search engine, has disrupted the advertisement-driven business model for newspapers or how Apple sells more watches than centuries-old Swiss watches.
They also discourage the creation of new companies by buying up all new potential competitors. This inhibits innovation.
Amazon Marketplace is one of the platforms that entrepreneurs are increasingly competing on. They may be at the mercy of abuses like fake product reviews from competitors or by unpredictable and opaque gatekeeper rulings. There is also the phenomenon of Digital Subsistence Entrepreneurs, online sellers who barely make a living wage.
This (anti-competition) landscape, sometimes called ” Platform Capitalism,” has caused regulators considerable headaches. Recently, the EU adopted proposals for a Digital Markets Act and a Digital Services Act to try and rein in actual and possible abuses of large AI-based digital platforms.
AI and automation would have been the driving force behind a rapid rise in labor productivity and a reduction in unemployment. We are instead seeing stagnant productivity growth. For example, in the UK, this is the lowest for 200 years.
Ireland’s AI Strategy ignores these problems of platform capitalism. Amazon and Facebook are not mentioned at all. Only Google is mentioned in the document. The document makes no mention of digital platforms, platform capitalism, the DMA, DSA, or the EU’s numerous antitrust actions. It’s like Hamlet without Prince.
Ireland’s AI Strategy should have stated how and when AI would achieve the economic benefits that it mentions, and who would reap them. The AI strategy should also have provided a vision for how the nation can avoid being a victim of the GAFAAs, or a mere agent.
Firms don’t use it
The strategy assumes, too, that people don’t understand AI well enough. Voila, teaching data science to people and appointing an AI ambassador (like a modern day prophet) is the solution. You might expect the exact opposite: the more people learn about AI, the less trust they have in it.
It is a good thing. AI adoption is low in the US, where AI understanding is advanced. In a recent US Census Bureau survey that included more than 800 000 US companies, only 2.9% of them were using machine-learning as recently as 2018. European Commission 2020 also found very low adoption rates.