It is as tempting as it is daring to make predictions about how Artificial Intelligence will impact the country's productive sectors. But the stake cannot be alien to the analyst, and especially to CAPEC, the Economic Policy Advisory Council of Catalonia, due to the positioning to be taken by the Generalitat. For this reason, in the last session there were several presentations and a debate from which I derived a set of reflections that I comment on here as 'food for thought'. The impact analyzed ranges from the most generative to the predictive, to simplify. And both for data learning topics (trained – machine – such as fondas or deep) – fully automated. However, ethical and regulatory considerations are left aside here.
The impact on productivity, the labor market and social inequality is of particular concern. A substitution effect on a certain type of labor force may make some of the immigration that dominates today more unnecessary than ever. With the alternative of AI, the need for these jobs may be reduced, or their wages may be pressured downwards. It is also likely that the technological requirement will leave less-trained workers out of the game, and will even put pressure on early retirement for those who believe that, at a certain age, they will no longer be able to achieve the required human capital formation.
It can also affect cognitive tasks that are relatively well paid today. Although AI generates a general positive externality on work (a kind of favorable impact, as the Internet had in what was the stage of growth called 'the great moderation'), it is likely that voices will be raised against the lack of a effective democratization of new technology. It will therefore be necessary to make those who will suffer the most share in the global gains, whether through universal or comparable citizenship income. However, it is not clear how taxation will absorb the profit from those automated incomes; and perhaps they end up concentrated in the “mega-rich”, who have a better capacity for tax avoidance.
Financing, therefore, of the expenditure necessary to appease the increase in inequality would not be guaranteed. Finally, in view of how states are positioning themselves to make exclusive use of AI, the likelihood that the gains from innovation and its use will reach everyone is reduced, particularly if it is used as a weapon of any kind. military dominance and power.
There are doubts about how this will change the business model: a lot of data means network and scale economies. If the algorithm is the added value, how it is charged for it is uncertain, beyond the exclusive use that may end up being made of the technology. Nor do we know the degree of anti-competitive collusion in prices that it may allow, nor the organizational change it causes.
Some areas in which Artificial Intelligence can be especially disruptive are mobility-driverless cars and health and social services, not to mention the arts in general. A look at healthcare can introduce us to a new world of healthcare services led by AI. In healthcare, the technological frontier continues to expand. Predictive medicine, personalized treatments, and the stratification of therapeutic indications are the best-known new milestones. The oncological field is especially suitable due to the great information it generates in data and diagnostic techniques. But, in addition, innovation (with respect to AI) can mean the rupture of the current ecosystem, with a technology that enables the entry of new agents both in demand ('non-patients', or not considered 'treatable' yet) as on offer (the active and non-reactive professional, to the extent that he predicts). This facilitates the emergence of an insurance and prevention intermediary broker, in the form of an algorithm managed even from outside the health profession, which guides the flows of multimorbidity situations in access to the system. As Genís Roca illustrates very well, “we can go from the situation in which someone feels unwell and calls their doctor, to another in which the doctor calls someone who believes they are well when the data warns that they are about to feel unwell. ”. Quite a revolution, as Anna Schlegel, also a member of CAPEC, observes. With all this, a new typology of 'patients' is born, more predictable through data than through relational contacts, and new actors appear from other sectors, new knowledge, with different business models for sharing information and knowledge, and with new financing schemes ( such as to reimburse, for example, dynamic algorithms). This may generate the need for new public-private partnerships with different relevance compared to the current binomials, as in the case of the industry and the public buyer. Of course, as long as it does not end up capturing the regulator.
These are just a few flashes of the changes in which we may be immersed.
