This note is based on a talk by Claude Revel at the “Artificial intelligence: legal and economic challenges in the information society” conference held in Paris on 13 October 2022 and organised by the MBA / Master 2 in Business Law and Management alumni association of the University Paris II – Panthéon-Assas, in collaboration with FIDES, Dalloz, Mayer-Brown and FIDIMA.
Between companies but also between States, harnessing AI has become an economic, political, and even a geopolitical power game. Public and private competitive strategies are challenged by AI at all levels. And yet, to date, the omnipresence of AI does not seem to have been questioned by private and public decision makers. What promises does this instrument hold for the common and individual good? Is its extension to areas that are more sensitive for humans, such as decision making, advisable?
Artificial intelligence (AI) has many definitions and they are evolving.
Prof. Ludovic Dibiaggio et al., the authors of a policy paper for our think tank, SKEMA Publika1, define it thus: “AI paves the way for innovations previously out of reach; it is expected to be a general-purpose technology, just like the steam engine, electricity and electronics were in past industrial revolutions. In the space of just a few years, AI has spread into sectors as diverse as transport, telecommunications, healthcare, education, justice and safety.”
But a new technology of this magnitude automatically brings with it new and fierce competition.
AI encompasses the entire science-technology-industrial applications chain. Competition thus occurs on all of these levels: between companies but also between States, since harnessing AI has become an economic, political, and even a geopolitical power game. All States able to do so support their economic agents.
The competitive strategies will therefore be private and public.
AI Challenges the Competitive Strategies of the Private Sector on Three levels: Business Practices, Prices and Patents
Competition in Business Practices
The common element in these strategies is that AI-based software applications promise to provide much more advanced services that are completely new in their approach, thereby eliminating even more man-hours. Many machines and products that cost a great deal to install will be rendered obsolete and the requirements of specialists will set back those who did not get in first. This is really nothing new, where disruptive innovations are concerned, but these transformations will occur on a very large scale and at the very root of processes.
In the introduction, we listed some of the fields impacted. Agriculture is another, as are the so-called “soft” industries, such as advertising, communications, journalism, and even works of art since AI applications are able to create images from text input for example.
Marketing has also been struck at its heart, by the transformation of so-called consumer engagement2 through the use of AI and soon the metaverse to engage with consumers and as a loyalty-building strategy. Offer strategies are accelerated and simplified as a result.
Another, of course, is management itself, resulting in significant competition-related issues for employees and managers. As a small – though not so insignificant – anecdote, an AI-driven robot named Ms. Tang Yu has been appointed as the CEO of one of China’s biggest gaming companies, NetDragon Websoft. According to the company, “Ms. Tang Yu presents advantages that cannot be matched by humans: she can work 24 hours a day without pay.” Of course, she is just the “marketed” presentation of an algorithm, but she certainly grabs people’s attention.
Business models are impacted at their root; those who are the best and quickest at mastering them come out on top.
AI and Price Setting
AI also has an impact on price setting. A 2019 report by Germany’s Federal Cartel Office focused on “Algorithms and Competition”. The Court of Cassation, the highest court in the French judicial system, discussed these same topics in a report dated August 2021. The issue is that pricing algorithms can help businesses to set prices but also to collude on these prices. The key challenge then becomes determining anti-competitive intent. In the beginning, that is relatively easy. But once they are trained, can AI algorithms collude with each other without humans even realising it? Can they have intent? This could be the perfect crime without intent. It is an interesting question for legal experts.
Patents3
It is through patents that AI is likely to have the greatest impact on competitive strategies globally.
Today, IBM is the patent leader with nearly 16,000 AI-related patents. Next come Intel, then Samsung followed by Microsoft and Japan’s NEC. Looking more generally at the top 20 biggest players in the field of AI, there are five American companies (the three mentioned above plus Google and Ford), two Korean (Samsung and LG Group), two Chinese (State Grid Corporation of China and Huawei), one German (Siemens) and one Dutch (Philips). The other nine players in the top 20 are all Japanese.
This ranking prefigures new competition, particularly from Japanese firms, tied to the exploitation of these patents.
For the record, the European AI leader, Siemens, is ranked 10th worldwide, with less than half the number of patents IBM holds. Thales, the leading French company by number of patents, is ranked 37th worldwide, with close to 3,000 patents.
This already indicates a radical shift in competition globally, still barely discernible among the Big Five. Microsoft and Google appear in this top 20, but Apple is right down in 27th place, while Amazon and Facebook rank 42nd and 49th respectively. However, this observation does not give the full picture for two reasons: on the one hand, companies can make the strategic choice to invest only in a specific field rather than all of them, and on the other hand the number of patents is only a partial indicator of the innovative capacity of firms, which may be investing abundantly in AI in other ways.
Legend: number of patents (blue) | number of inventions (purple).
Source: Artificial Intelligence: a Political Subject.
Public Competitive Strategies: Soft law, Hard law, and Diplomacy
Harnessing AI is too strategic for States not to get involved. Their interventions will of course influence competition.
According to the Tortoise Global AI Index 2021, which benchmarks nations on their level of investment, innovation and governmental implementation of artificial intelligence, the United States and China are in the top two spots. France and Germany are in 9th and 10th place respectively, right behind the Netherlands.
At least three fields compose public competitive strategies: norms, the law, and diplomacy.
The Battle of Standards
Whoever sets the standards sets the playing field. This issue involves companies but also the States, which traditionally play a more or less active role in this normative competition.
At the European Union level, at a press conference on the EU Standardisation Strategy held in Brussels on 2 February 2022, Thierry Breton, the Internal Market Commissioner said: “Technical standards are of strategic importance. Europe’s technological sovereignty, ability to reduce dependencies and protection of EU values will rely on our ability to be a global standard-setter.” In fact, he largely cited AI as an example.
China is very active in this domain. Conscious of the competitive edge brought by standards, Beijing is channelling more and more efforts into setting the international standards for AI, as detailed in the China Standards 2035 strategy. In fact, in 2017 China announced a $150 billion plan to become an AI superpower by 2030.
In France, there is an Investments for the Future programme (PIA or programme d’investissement d’avenir) called the Grand Défi Intelligence Artificielle (the great AI challenge). According to its director: “We are faced with a major industrial challenge, with a market estimated at 50 billion euros.Supporting the ecosystem requires standardisation and establishing a solution based on trust.” AFNOR (the French Standards Association) and players in the industrial sector are now participating in the technical subcommittee ISO/IEC JTC 1/SC 42 Artificial Intelligence and, in the EU, in the work of the CEN-CENELEC Joint Technical Committee on Artificial Intelligence (CEN/CLC/JTC 21).
The Widening Gaps in Competition Law
The importance granted to AI by the States is leading to significant changes in long-established legal concepts of competition. A few examples are listed below.
In its March 2022 report, on the subject of public procurement, the French Council of State wrote the following: “When an artificial intelligence system interferes with the fundamental interests of the nation, it would seem justified that bidding be restricted to companies entirely bound by European law,” due to the risk of abusive data harvesting practices. Also, that the public AI strategy must be designed in such a way as “to safeguard the sovereignty of France and guarantee the strategic autonomy of the nation.”
In addition to its gigantic public-private R&D efforts, the United States government intervenes in private-private business. As an example, Nvidia Corporation announced it had received a letter from the US government on 31 August 2022, banning them from exporting advanced AI chips to China. AMD reports the same thing.
Finally, a New Issue Will Emerge: AI as a Competitive Tool in State Functions
AI is increasingly being used in core State functions. Already, in France, it would seem that the Cassation Court is going to use AI to help verify the consistency of jurisprudence nationwide. Other countries, such as Estonia, have already gone much further than this, by using AI to prepare or even rule on cases. The field of justice is not the only one concerned. It is very likely that there will soon be rankings – and thus a new form of competition – to determine which States “best” use AI to govern; the question then becomes how to define “best”, somewhat like the Doing Business ranking intended to help assess regulatory performance.
Lastly, in the field of diplomacy, AI can become a ‘soft power’ tool, particularly through the provision of technical assistance and training in developing countries. Certain States, or private players working with governments, deliver training programmes to less developed nations – obviously for some advantage. The United States does this via foundations and universities. For example, Carnegie Mellon University and the Mastercard Foundation have partnered with the Rwandan government to develop engineering and technology programmes. These entities fund the creation of a degree in AI engineering, among other programmes. The rationale behind this is that by 2030 there will be hundreds of millions of youths in the African labour market and they should be trained, particularly in AI, according to American standards and practices.
Why Artificial Intelligence?
It seems obvious that AI is going to transform the competitive strategies of the private and public sectors globally. The topic of AI is everywhere and it would seem that no sector can escape it. And yet, to date, its omnipresence does not seem to have been questioned by private and public decision makers. Beyond the questions of ethics, privacy and democracy raised by this new “universal instrument”, what hopes have humans placed in it? What promises does it hold for the common and individual good?
While AI does make new and useful applications possible, such as improved medical diagnosis, driver assistance technology in vehicles, smarter resource allocation in agriculture, help with creation itself, and easier access to information, is its extension to areas that are more sensitive for humans, such as decision making, advisable? Does its use in the fields of recruitment or justice truly bring added value? Has this been evaluated?
It seems essential today that policy makers perform a proper ontological analysis of algorithms and their impact before deciding on vast programmes to develop AI in all sectors. In short, might AI be yet another magic word used by private and public decision makers wanting to stay in step with their times? Might we be experiencing an umpteenth temptation to turn to the “techno-solutionism” so familiar from past industrial revolutions?
1 With the researchers Lionel Nesta and Mohamed Keita, also from SKEMA Business School.
2 Research topic of Prof. Margherita Pagani, SKEMA Business School.
3 Data taken from the SKEMA PUBLIKA policy paper on AI (2022).
