Expertise has a tensed relationship with democratic societies. Western philosophy was arguably born in opposition to democratic Greek poleis as an effort to rescue the role of the experts from the pretenses of the crowds, the majorities of ‘lay’ people who dissented to the rule of the knowledgeable few. Such tensed relationship between expertise and democracy explains why certain features of the internet have been widely praised as having the potential to deepen democracy. In particular, the replacement the experts with the crowd is one of the most important democratic promises of online networks.
How could the internet replace expertise with the crowd? There are fundamentally two distinct mechanisms here: (1) crowd mining and (2) peer-to-peer production.
Crowd mining. Learning algorithms could track and mine the online behavior of millions of people to produce meaningful knowledge. Google knowledge graph is the archetypal of such algorithms. By tracking the ‘clicking’ behavior of users of google web services, the algorithm is able to match search queries with their most relevant answers. It does so by quantifying the time that users spend on each ‘results’ page after a google search. If the user spends a lot of time on a specific page, it is likely that the page was highly relevant for her search. If instead the user immediately ‘click’ again, in all likelihood the page was poorly related with the original query (ref 2). Automatic Google results, i.e. those that instantaneously pop up without clicking (also used in certain apps), are based on such algorithm.
Peer-to-peer. The paradigmatic tool for peer-to-peer production are wikis, which allow the joint collaboration of several people on texts, scripts, etc. While sometimes these tools may contain reputation-based restrictions on contributions, thereby reproducing the form of expertocratic organization of many offline research endeavors, they make the production of knowledge much more horizontal. In particular, in wiki-like production of knowledge, reputation does not depend formally on certificates, colleges, references. The online encyclopedia Wikipedia is arguably the most successful wiki-based intellectual product, and famously Giles (ref 3) reported in Nature that its accuracy matches that of expert-based Encyclopedia Britannica.
There is a number of ways in which the democratic promise of the internet may fail. Many have argued that online patterns of monopoly on the production of knowledge are very similar to their offline counterparts (e.g. newspaper, TV), with few hubs receiving most attention. Other analysts ask to pay attention to the distribution of power in those forms of knowledge production that depend on proprietary web 2.0 platforms (ref 1), and to the wealth inequalities associated with internet (some key players of the internet are among the richest people in the world).
Moreover, the replacement of experts with the crowd is a phenomenon that is much more ambiguous than the positively connotated term ‘democratization’ may suggest. This could be seen if we reflect on the limitations of such replacement.
Are there field of expertise that are safe from replacement with the crowd? Basic science, both natural sciences and the humanities, may not lend itself to replacement with crowd-based forms of knowledge production. But the effects of internet on our everyday relationships with experts may be profound. Certainly many of the readers already ask google or wikipedia when they are told that an acquaintance has a certain disease, or even when they start experiencing unpleasant symptoms. This is at any rate the first strategy for many, at least because a PC or a smartphone are most often more close-at-hand than doctors. Here experts are indeed losing some ground.
Despite these defeats, experts may have recently scored an important goal in their game with the crowds. Google is now partnering with the US Mayo Clinic to curate health related information that gets displayed on the basis of google knowledge graph. Physicians will double check for accuracy any Google health information that appears in instantaneous search results (ref 4). The accuracy of health related searches is a particularly sensitive topic. Both legal and moral standards are stricter in the case of health information, as opposed to other domains (cf. gardening, or birdwatching), because of the potential severe harms that can result from the release of misleading information.
The partnership between Google and the Mayo clinic is a further widening of google-based services. However, it is also a neat defeat for the promise of democratization illustrated by google algorithms as knowledge graph. In fact, the collaboration with Mayo clinic can be seen as an admission that experts are still superior to algorithms when it comes to the production of sensitive health information.
Moreover, such partnership suggests caution in referring to crowdsourced models of knowledge production as instances of democratization. Google will employ on average 11 experts to double check each health related query. We do not know the terms of employment but we may imagine it will be waged employment (at least indirectly). This reveals the ambiguity of forms of ‘democratisation’ based on crowdsourcing. Crowds are unpaid. This fact may be not unrelated to the emergence of wealth inequalities associated with the internet. Much caution is necessary in the evaluation of this empirically controversial causal claim. The emergence of crowd-based forms of production may ultimately play a very small role (or none) in the historically very high level of wealth inequalities of our contemporary world. Even if some entrepreneurs relying on crowdsourcing are indeed very rich, and some traditional intellectual jobs are losing terrain, other factors are at play. However, and independently from this open empirical question, democratic enthusiasms towards the replacement of experts with crowds ought to be tempered with a balanced assessment of possible distributive consequences of the (partial) demise of professional experts.
Kostakis & Bauwens (2014) Network societies and Future Scenarios for a Collaborative Economy. Palgrave. Cf extract here.
Adams (2013) Google and the Future of Search. The Guardian.
Giles (2005) Internet Encyclopaedias go head to head. Nature.
Gibbs (2015) Google to put health information directly into search results. The Guardian.