My view of the digital world has been changing noticeably in recent years. I may also look at it differently from someone who has only recently become involved with digitalisation. As an elder person, I have personally witnessed much of the development of the digital world since 1965: the emergence of computers, the internet, mobile communications and social media. Now, I am witnessing another fundamental transformation, driven by the emergence of large language models (LLMs) and artificial intelligence in general.
At the same time, it is not only technology that is changing, but also the world in which that technology is developed and applied. Global power relations are shifting. For Europe, this means, among other things, that we need to reconsider our relationship with the United States. For a long time, we took it for granted that the US was part of the same Western world and a natural partner. That view can no longer be maintained without question. At the same time, it is becoming increasingly clear that cooperation within Europe is essential. China's rise is also encouraging Russia to adopt a more aggressive stance towards Europe. These developments reinforce the need for cooperation and compel Europe to rethink its own position and security.
In my view, all of this calls for a different perspective on the digital world, and particularly a different way of looking at knowledge and innovation. Knowledge is no longer merely an economic or scientific resource; protecting knowledge has also become a matter of military and strategic importance.
In fact, this has been going on for much longer. It began with the acquisition of knowledge and information. In the 1990s, for example, Microsoft purchased large quantities of digital content, including images of works of art and encyclopaedias, thereby building a significant knowledge position. Later, companies such as Google offered citizens a range of free services in exchange for access to vast amounts of personal information.
As a result, American technology companies now hold an irresponsibly large amount of data about European citizens. This raises questions not only about privacy but also about strategic dependence. Through their software and ecosystems, companies such as Microsoft and Google have also managed to create strong lock-in effects. Many organisations and citizens have consequently become heavily dependent on American technology for their day-to-day activities. I regard this as a major threat.
The problem is that AI brings fundamental questions into sharp focus—questions that we have been unwilling to confront for several years.
Many articles currently discuss the inherent risks of today's LLMs: systems becoming more intelligent than humans, the possibility of developing extremely dangerous viruses, or models that begin to exhibit criminal behaviour through recursive development. My focus lies elsewhere. The quantum leap in capabilities offered by current AI makes existing risks in the digital world more apparent, including the risk of attacks on our society. The recent attack on Albert Heijn is one example.
The role of human beings is also changing because of the rise of AI. On the one hand, we see AI taking over certain human decisions, putting concepts such as liability and accountability under pressure. On the other hand, people can easily hide behind their anonymity in the digital world. Anonymous threats against politicians are reaching levels that, until now, we associated only with the criminal underworld.
The question is how organisations should deal with knowledge and innovation. In my view, there are roughly three possibilities.
The first is simply to keep knowledge secret. If you have come up with something clever, you can decide not to share it with others. The classic example is Coca-Cola, which keeps its product's recipe secret. In this case, knowledge is valuable precisely because others cannot access it.
A second option is to protect knowledge legally. You prevent others from using it without restriction by, for example, applying for a patent or claiming copyright. Knowledge thus becomes a protected asset whose use is subject to certain conditions.
The third option is to decide that sharing knowledge contributes to your own development and business. This can be done by publishing white papers or developing and supporting public standards. By making knowledge available, an organisation can influence the further development of a technology or market. Openness then becomes a strategic choice.
There is another dimension to consider. People can also develop and share knowledge for reasons that are not directly economic. Creating knowledge can give people a sense of purpose and the feeling that they are contributing to society. Open-source software is a good example: people invest time and expertise in something that does not necessarily offer an immediate financial return. Wikipedia is another example of this model. In fact, sharing knowledge has been a cornerstone of science for centuries, enabling progress because researchers build on one another's work.
Throughout all of this, it is important to recognise the differences in interests among the various stakeholders. Businesses focus primarily on profitability, both in the short and the long term. Governments have a very different purpose: they serve the public interest and focus on security and the economy. Individual citizens are concerned with privacy and a sense of purpose, alongside their primary responsibility for themselves and their loved ones. Healthcare institutions occupy an intermediate position: in our society, they are partly businesses, but they are also part of the public sector. Each of these stakeholders has a distinct role in the world of knowledge and innovation.
I became familiar with concepts such as open source and open systems between 1980 and 2000, both through my work at the university (where Bell Labs made the Unix code publicly available) and through collaborative initiatives involving telecommunications companies, or for the whole ICT industry via X/Open. As technical director of one of these collaborative telecommunication organisations, I had my office in Reading, in the same building as X/Open.
A defining characteristic of these projects was that all parties recognised that sharing knowledge benefited society, but also, and especially, their own businesses. Keeping certain knowledge within one's own organisation did not make sense. Eventually, even IBM, which was dominant at the time, realised that proprietary standards were counterproductive.
The big question now is whether an open-source approach can survive the current, extremely rapid development of AI and the significant military and strategic interests involved. Nevertheless, smaller players, particularly smaller governments, should consider pursuing an open-source approach.
During the conference, the question of how we should approach open source came up. I found this an interesting question, one that is becoming even more relevant as AI advances. At present, it is still difficult to predict how society will ultimately change as a result of AI. We are in the middle of a development whose consequences are far from clear. The debate about open source will only reach a conclusion when there is a broadly shared understanding of how we want to deal with knowledge in the future.
Perhaps, therefore, we cannot separate this debate from a much more fundamental question: how do we, as a society, want to deal with knowledge in the future? Do we primarily want to protect knowledge because it represents power and strategic value? Do we want to exploit knowledge commercially and restrict access to it through legal protection? Or do we see the sharing of knowledge as an essential condition for innovation and social progress?
My impression after attending HHHAI is that the debate about AI is ultimately about much more than technology. It is also about how we organise our society and the value we attach to knowledge. Only once we have developed a broadly shared understanding of these issues as a society will we be able to determine how we want to approach AI, open source and innovation.
Below, I formulate a few propositions about possible changes in the digital world.
The fundamental view of the internet today is that of a worldwide network. Now that the dividing lines between geopolitical power blocs are becoming so dominant, a more hierarchical structure for the internet would be more appropriate, with strict controls to ensure that certain data is not transmitted across borders. Personal data, for example, should not be sent outside Europe. I find it strange that all kinds of data about my household are stored in the cloud without any filtering.
Anonymity in the digital world is regarded as a great good. It is understandable that certain groups have a need for anonymity, but the way it is safeguarded needs to change. In journalism, this is well regulated: journalists have the right to protect their sources and are not required to disclose all of them. Comparable rules apply to priests and doctors. The rules governing anonymity in the digital world also need to be developed more thoroughly.
One lesson from this period is that almost all the original companies that launched the computer industry have disappeared. Unlike the oil industry, the digital world is far more transient. Even data becomes outdated quickly, as I learned from the PTT's 008 database. The automotive industry has also lost many of its early winners, as illustrated by the British car industry.
I also saw first-hand how computer companies such as Microsoft and IBM operated in this environment and chose to pursue standardisation.
A second lesson is that the second mover can be highly successful. Both Apple and Microsoft are, in effect, second movers. Apple built on the shoulders of Xerox. Microsoft was guided by IBM, which recognised a market need for the personal computer; in doing so, IBM looked, among other things, at the gaming industry and the process and laboratory industries. Apple itself began in the design studios of marketing companies, while the IBM PC only became a success thanks to accountants using what was a precursor to Excel. MS Office emerged in response to a request from British Telecom.
The AI revolution is making the problems in the digital world increasingly visible, and these problems are creating an ever more urgent need for solutions.
