The diffusion of AI is making possible innovations that are reshaping our daily life by augmenting and complementing human capacities in various domains, among others: self-driving vehicles, image processing and clinical diagnosis. For some, this is the dawn of the 4th industrial revolution, which is expected to entail deep transformations in the production and consumption processes. Such transformations could lead to significant productivity gains. Accenture (2016) recently estimated that in 12 advanced economies AI could double annual economic growth rates by 2035 and boost labour productivity up to 40%. However, there is no agreement that accelerating the adoption of new technologies will reverse the productivity slowdown experienced since the mid-2000s in mature economies as well as in emerging markets. The increasing digitalisation of our societies and economies has not yet been reflected in increases in productivity (as in the “Solow paradox”). Different explanations have been proposed to capture the productivity trends. The first explanation highlights the limits of current statistics in fully capturing the benefits of new technologies. To better reflect the ongoing transformations, definitions and measures should be adapted at both national and international level. An alternative explanation is that the existence of barriers prevents firms from embracing new digital technologies, including AI solutions. Policy and structural factors lifting firms' capabilities (e.g. organisational capital, ICT skills and efficient allocation of human capital) and sharpening market incentives to adopt new technologies (e.g. labour market flexibility, competitive pressures and risk capital availability) could play a relevant role in supporting the diffusion of AI. Finally, potential impacts of new technologies take time to emerge at macroeconomic level. This partly depends on the time it takes to build up a stock large enough to impact at macro level and partly on the need to realise complementary investments to reap the full benefits of new technologies. Positive impacts on productivity might therefore materialise by moving up along the process of deployment of AI technologies. Public institutions can play a relevant role in fostering the productivity gains deriving from the diffusion of AI. They can stimulate investment in research and deployment of AI and remove barriers for firms to adopt new technologies, including the promotion of an appropriate education and training system offering life-long learning opportunities. The scale of needed investments and their positive spill-over effects justifies intervention at EU level. In line with the Communication adopted in April 2018, the Commission is increasing its investment under the Horizon 2020 research and innovation programme to €1.5 billion for the period 2018-2020. Additionally, the European Fund for Strategic Investments is expected to mobilise more than €500 million in total investments by 2020 to support companies and start-ups in investing in AI across a range of key sectors. Proposals under the EU's next multiannual financial framework (2021-2027) include strengthened support for training in advanced digital skills, including AI-specific expertise.