Will AI mitigate or exacerbate global instability?
When it comes to capitalizing on artificial intelligence and data analytics, some countries and companies are ahead of others. This was the view of some of the speakers attending an online Science | Business workshop exploring the Great global data divide.
Describing the ‘disturbing’ network effects that concentrate artificial intelligence (AI) capabilities in the hands of a few, Nicolas van Zeebroeck, professor of digital economics and strategy at the Free University of Brussels, said: “It’s pretty clear that the imbalances that you see growing in terms of profit accumulation, but also in terms of skills and data concentration, for example, are a potential source of instability at all levels. In the same spirit, research by the International Monetary Fund warned of “a big divergence” caused by AI and robotics (see table).
For van Zeebroeck, the “two animals” that policymakers must tame are “insufficient professional and social mobility within and between countries, and growing barriers to entry and innovation due to the accumulation of data and skills ”in a few hands. He noted that developing effective AI requires a combination of the right data, the right skills, the right culture, the right infrastructure and the right software. “You need the whole ecosystem in place and this is what Europe and the developing countries need to invest in,” he added.
But how much should they invest? While it is clear that countries and companies need to enter data and implement it, it is very difficult to put a monetary value on this activity. “Many economists, including those at the OECD, are trying to figure out how to value [data] and how to understand its value, ”noted Audrey Plonk, Head of the Digital Economy Policy Division, OECD Science, Technology and Innovation Directorate. “And that’s a tricky, tricky puzzle.”
In terms of patent applications and research collaboration, “we see the US and China way ahead of many other countries, including the EU, if you take them together,” Plonk added. “We are publishing new data in the coming weeks on [venture capital] investments, and in that you will see that the United States is ahead of China and everywhere else. However, she warned that such measures only tell part of the story and do not necessarily imply a widening gap.
In addition to public sector investments in research and development and skills, a coherent national strategy can be important. “We have seen a proliferation of national AI strategies over the past two years,” Plonk said. “The minute the government sets high-level goals […] it inspires a whole host of activities, both in the private sector and in the public sector. ”
Count the cost of connectivity
For countries in the South, one of the main obstacles could be the cost of digital infrastructure. There is a “data divide based simply on the lack of affordable connectivity,” explained Cathrin Stöver, communications director of GÉANT, which connects European scientists with each other and with their counterparts elsewhere in the world. “And I must point out that it is no longer a lack of cables. It used to be, but it is no longer the case, ”she added.
Although Africa is one of the best connected continents in the world (in terms of cables serving it), bandwidth between African cities can cost € 60 per megabit per month, compared to just two euro cents. between European cities, according to GÉANT. “In some parts of Africa, connectivity is not just a bit more expensive. It’s 3,000 times more expensive, ”noted Stöver. “And this, from our perspective, is one of the main drivers of the continuing digital divide.”
During the pandemic, the lack of affordable connectivity led to the complete shutdown of education in much of the south of the world, Stöver noted. “It is absolutely terrible for the development of these economies. And I think that shocked a lot of stakeholders. To help GÉANT lower the cost of connectivity, Stöver called on the European Commission to establish more programs including regulatory dialogue, alongside infrastructure investments. At the same time, more activities need to be “driven by the South” to find solutions that work and not just “copy and paste activities from Europe or the United States because they work in our economies. She added.
Towards a fairer economy
A growing gap between the “data haves” and the “data haves” is also apparent in different sectors of the economy. Describing the results of a study based on 50 in-depth interviews with organizations in Europe, Atia Cortés, post-doctoral researcher at the Barcelona Supercomputing Center, said that there is “a different maturity in the levels of AI being implemented. and integrated into the business models and this was independent of the business sector. ”She pointed out“ a big difference ”between large companies, SMEs and research centers reflecting differences in their financial and human resources.
Cortés called on Europe to develop strategies to boost AI adoption “at all levels”, while putting in place mechanisms to ensure that local communities get more value from the data they generate. “The main challenge facing governments, especially here in Europe, is how to retain the talent created here, both in academia and in the private sector, especially in SMEs, which is one of the foundations of our economy, ”added Cortés. .
Indeed, the vital role of skills and education was a recurring theme of the webinar. As AI permeates so many aspects of daily life, it may be necessary to ensure that anyone in authority (regardless of their area of expertise) understands AI. Teresa Scantamburlo from the European Center for Living Technologies, Ca ‘Foscari University in Venice, highlighted how several northern European countries have introduced AI into a wide range of university courses, including humanities, while expanding the courses computer engineering to take into account the ethics and societal implications of AI.
While Scantamburlo argued that governments can play a central role in ensuring that AI is ethical and trustworthy, she also stressed the need to “Have other independent organizations that review the work that governments do”. In particular, Scantamburlo pointed to “the specter of increasing scrutiny as well as the opacity of bureaucracy and public services” as issues that need to be monitored by independent organizations.
Can Europe influence the rest of the world?
A key pillar of the EU’s digital agenda is developing a ‘trustworthy’ approach to AI, distinct from the freer approach taken by the US and the state-centric strategy pursued by China. At the same time, the EU is encouraging greater data sharing both in Europe and globally.
Microsoft, which has been running a global campaign on open data for 18 months, sees the EU strategy broadly in line with its view that wider dissemination of data is beneficial. While welcoming the new EU Common Data Spaces (initiatives to pool public and private sector data for mutual benefit), Jeremy Rollison, Senior Director of EU Government Affairs at Microsoft, highlighted the need to find the right balance between advancing European companies and research and ensuring they have access to the best tools. “What direction are we going to take in terms of intellectual property in this space? ” he added. “How do you create a framework that always encourages data collection, data retention and promotes access to that data?” “
Rollison stressed that Europe needs to carefully measure the impact of its initiatives in terms of “how exact data sharing [they] create. “If successful, the evolving regulatory framework for AI and data in Europe could have a big influence on third countries, as could the General Data Protection Regulation (GDPR) of the EU has been widely emulated. “If data sharing takes off and competitors in Europe are more comfortable participating in joint public-private initiatives or sharing data between them that boost their global competitiveness, then I think you will see the rest of the world emulate that as much as possible, ”he added.
Of course, there is also a risk that overly rigid regulations slow down Europe. The EU must ensure that the rules allow organizations to access the best talent, tools, technologies and platforms, Rollison stressed. “Openness is a key word here, both in the context of data, and in the context of accessing talent, technologies, types of software. ”
Winning hearts and minds
However, not everyone is comfortable with the concept of open data. Hilary Hanahoe, secretary general of the Research Data Alliance (RDA), acknowledged the “fear people have around the world open”. She stressed that “open data is not necessarily free data”. Building on comments from other speakers, Hanahoe underscored the need for trust, as well as skills and standards, to bridge the global data divide.
Noting the need to build trust in a global ecosystem for sharing data, as well as between people and technologies, she underscored the importance of ensuring that people and communities in developing countries remain in control. of their data. To this end, the RDA and the Global Indigenous Data Alliance have co-developed the CARE Principles to guide data management. Designed to be complementary to the widely adopted FAIR principles (findable, accessible, interoperable and reusable), the CARE principles call for “collective benefit, supervisory authority, accountability and ethics,” Hanahoe explained.
She described CARE’s principles “as very, very relevant to many global challenges, [such as] climate, population health and urban health. Indeed, several workshop participants noted that the pandemic underscored the need for greater global cooperation and data sharing to address the many threats that cross international borders, such as infectious diseases, climate change and demographic changes.
While calling for action to bridge the digital divide, Hanahoe stressed that policymakers should not lose sight of the fact that “a digital revolution is the bright side of this – there are a lot of opportunities”.