Lucubrate Magazine Issue 38, Friday 7th, September 2018
An ongoing conversation about artificial intelligence and the key governance and innovation issues facing policymakers, scientists, innovators, industries and academics, powered by our founding partner Accenture. POLITICO hosted an invitation-only roundtable discussion, in partnership with Accenture Applied Intelligence, in Paris on June 14, 2018, to brainstorm and suggest pragmatic solutions for using artificial intelligence to serve French citizens. Here are the main takeaways:
1. The battle for artificial intelligence: A battle for human intelligence
Just a few years ago, the term was in the realm of science fiction. Artificial intelligence is now giving rise to transformation every day, and all the stakeholders have now realized the decisive role AI will increasingly play in the future. Data collection and management, virtual correspondents, process automation, business intelligence: AI has become a concrete reality. It is radically changing whole sectors of economic activity. It requires a reappraisal of legal concepts which previously seemed very sound. AI will to a certain extent exercise control over our lives. But who will exercise control over AI? Will we succeed in using it at the service of citizens?
These issues have come into the limelight in France in recent months. The Villani report and an artificial intelligence plan express the government ’s determination to have France play a leading role in this matter. And the European General Data Protection Regulation governing the use of personal data came into effect in May. Beyond merely the competition and regulation aspects, these issues also can lead to the emergence of new models of partnership and exemplarity.
2. A transformation already underway
The French higher education system has conducted pioneering research for a long time. The administration and large local and regional governments have followed a path over the past few years that has led them from using data science to real artificial intelligence tools. The state, aware of its lag and its inertia, has in the past 18 months taken a measure of these issues and shown a proactive approach to reform.
The impact of AI is also being felt in the private sector. As one speaker said: “Artificial intelligence is radically transforming companies. It’s a change that had not been seen in the past 30 years. It’s an absolutely unique change.”
This change is admittedly not an easy one. Many people are likely to be left behind if they do not receive guidance in upgrading their skills and expertise. The various actors in the public sector, who are each working on a specific aspect of artificial intelligence, were originally skeptical of the idea of working together, or with German or European partners.
However, the initiatives are starting to bear fruit and have sometimes laid the groundwork for new partnerships in which government agencies, higher education entities, and local and regional governments work on projects jointly with big companies or startups. Conversely, it is sometimes the startup spirit that impregnates traditional institutions. One participant explained the path followed on many projects as follows: “We act, we progress, we demonstrate, and we encourage others to follow.”
Transformation. Picture: Pixabay
3. Multi-dimensional data
The fuel of AI is data. On this, the participants pointed out the advantage of the French system, with its highly centralized government departments. It, therefore, has databases of gigantic proportions, notably in the area of health, which can then be exploited.
However, this data cannot always be put to use immediately. First, there is the issue of the data’s quality, format, and standards. The issue is crucial for decision-making tools, but even more so to share information across separate databases.
Now, as one participant observed, sharing is not obvious: “Data is a symbol of power. Throughout history, it has been the property of the people who held it.”
In the case of government departments, it should be remembered that this data is above all the data of the state rather than of a particular department. And big companies want to keep control of data generated by autonomous vehicles.
These obstacles must be overcome. However, the participants agreed in considering that a central agency in charge of data collection would present a number of problems. Data collection and management must, of course, protect citizens and their personal information. But GDPR also aims to facilitate the responsible reuse of data and to make it an opportunity rather than an obstacle.
Finally, social networks and cloud computing services have a considerable volume of data to monitor and supervise, which is why AI systems are proving necessary. They are still perfectible, and the use of human judgment is essential. “Artificial intelligence is not yet very familiar with subtle humor,” noted one guest, who stressed the need to defend humor and parody, which machines find hard to detect.
4. The strengths of France and Europe
The participants pointed to the fact that the government was a relative latecomer on AI, but that it is prepared to devote the resources to catch up. France has a role to play, notably in Europe.
As one speaker said, this is a strategic issue: “There is a European model to be designed, between on the one hand an American model, where AI policy is driven by large private players, and a Chinese model, closely controlled by the government, with a debatable approach from the ethical and democratic perspectives.”
France can already highlight the quality of its higher education system and its fundamental research (including on creative AI systems), together with tax benefits. It is attracting operations set up by these large private players and also by smaller organizations. Moreover, a dynamic seems to be emerging, fueled by the investment plan and even by Brexit, to anchor startups in France and put an end to the brain drain.
This is a fundamental aspect: “The battle for artificial intelligence is above all a battle for human intelligence,” noted one guest. “We won’t overcome this problem without having a research ecosystem capable of welcoming the best scientists in the world.”
5. The big foreign private players: Best enemies
The large foreign private players specialized in AI are not just limited to the GAFA (Google, Apple, Facebook, and Amazon). They are perceived as both allies and threats. They set up research laboratories in France and share their work, and have become aware of the fight against cybercrime and illegal content.
But they are buying up European startups, sometimes taking over their intellectual property in AI research. Their economic weight alone gives them huge clout and creates imbalances.
The autonomous vehicle industry plans to negotiate separately with each major city and retain full ownership of data. In response, these cities have united to define a common position. Some robotics companies have pushed for their products, nurtured on AI, to obtain a legal personality, reducing their own liability in the event of an accident.
In conclusion, these players are partners to be watched closely. For one speaker, the European elections of 2019 will be “the chance to transform these ethical responsibilities, which these companies fundamentally recognize, into legal and legislative responsibilities at the European level.”
Lucubrate Magazine Issue 38, Friday 7th, September 2018
First published by POLITICO 7th August 2018
Illustration: PC Tech Magazine