An editorial by Jenny Ebermann, CEO Wikimedia CH
The importance of open, accessible online knowledge for democracy is undeniable. But what would happen if AI-powered tools such as ChatGPT increasingly replace the use of online encyclopaedias such as Wikipedia? This development would raise significant questions and challenges that could have far-reaching implications for the quality and availability of knowledge. The integration of AI tools into knowledge access undoubtedly holds potential and opportunities. AI can provide information quickly and in an easily understandable form, which facilitates access to knowledge. At the same time, however, measures must be taken to ensure the quality, transparency and diversity of information.
One of the biggest challenges is the quality and reliability of the information provided by AI tools. AI models such as ChatGPT are based on large amounts of data that may contain erroneous or biased information. Unlike referenced information, such as on Wikipedia, AI runs the risk of spreading misinformation. This can lead to a distortion of knowledge and misunderstandings that are difficult to correct. “As a society, we should ensure that AI models are increasingly trained to rely on credible sources, e.g.: academic journals, publicly funded studies and public service media. Of course, this requires a functioning economic model,” says Dimi Dimitrov, EU Policy Director of Wikimedia Europe in Brussels.
Another problem is the lack of transparency and traceability of the sources used by AI tools. It is often unclear where the information comes from, which makes it difficult for users to assess the credibility of the content provided. This can undermine trust in the information provided. On the other hand, it can encourage them to consume information uncritically. The increasing dependence on a few large (AI) providers can also limit the diversity of available information sources and strengthen monopolies. This could jeopardize the diversity of opinion and plurality in society, which is a key prerequisite for a functioning democracy. If only a small handful of companies control access to knowledge, this will considerably influence opinion by homogenizing the formation of beliefs and certainties.
The digital divide also poses another problem. Access to advanced AI tools often requires technical infrastructure and digital literacy, which could deepen the divide between different population groups. People who have no access to these technologies or who are unable to use them could be partially or totally excluded from access to knowledge, which can only reinforce existing inequalities.
The spread of misinformation through AI tools can lead to the polarisation of the masses and social division. This undermines democracy, as numerous current examples show. Responsible use of such technologies is therefore crucial to protect and promote democratic values. It is important that these developments are critically monitored and shaped by society, academia and regulatory authorities.
Another relevant question is how to deal with the fact that commercial companies use the work of volunteers for free by extracting and reusing data from wiki platforms. These companies profit from knowledge that is free and open to all without offering sufficient compensation. This poses a significant risk to the Wiki Movement, which relies on voluntary contributions and donations.
The dilemma of the Wiki Movement is that it wants to make knowledge available to everyone in an open and reusable form, including commercial use. However, restricting access conditions would also harm many smaller organisations and individual users, as Dimi Dimitrov points out. To address this problem, the Wiki Movement is trying to square the circle by keeping content open to all but trying to persuade large platforms and services to contribute a fair share financially. The idea is to charge a fee for a direct, fast connection to the servers and, if necessary, to provide customised APIs. Content can of course continue to be used, but if a company makes tens of thousands of requests in real time within a short period of time, this puts a strain on the server infrastructure and leads to bottlenecks, for which a paid solution then must be found. The initiative in question is called Wikimedia Enterprise. It is a first attempt to move in what is hopefully the right direction. So far, Google is a paying customer, but the Wikimedia Foundation has offered free access to the Internet Archive and DuckDuckGo, because diversity and competition are valued. The hope is that this will become one pillar among many for funding free knowledge. However, to keep the projects and associations worldwide independent, the Wikimedia Movement needs the continued support of as many donors as possible.
By using free knowledge platforms responsibly and supporting them, everyone can work together to ensure that open and accessible knowledge remains a cornerstone of our democratic society!
