According to a recent study, LinkedIn conducted experiments on more than 20 million users over the course of five years. These experiments were designed to improve the functionality of the platform for its members; however, the results of these experiments may have had an impact on the livelihoods of some users.
In a series of experiments that took place in different parts of the world between 2015 and 2019, Linkedin changed, at random, the ratio of weak contacts to strong contacts that were suggested by its “People You May Know” algorithm. This algorithm is the automated system that the company uses to suggest new connections to its users. Later on, researchers from LinkedIn, M.I.T., Stanford, and Harvard Business School reviewed the aggregate data from the tests in a paper that was published in the journal Science earlier this month.
Tech behemoths such as LinkedIn, which is the biggest professional network in the world, often conduct extensive experiments on a wide scale, in which they test out various iterations of app features, site designs, and algorithms on a variety of individuals. This time-honored method, which is known as A/B testing, is used to enhance the customer experience and maintain customer engagement. This, in turn, enables businesses to generate revenue by charging premium membership fees or selling advertising space. Users are often unaware that corporations are doing research on them in some capacity.
However, the adjustments that LinkedIn implemented are an example of how seemingly innocuous tweaks to algorithms that are used by a large number of people may turn become social engineering experiments that may have life-changing repercussions for a large number of individuals. The conduct of long-term, large-scale tests on individuals that may alter their career chances in ways that are unseen to them sparked worries about the industry’s openness and research supervision, according to experts who study the social consequences of computers.
An influential sociological theory known as “the strength of weak ties” was put to the test in a study that was published in the journal Science. This theory maintains that individuals are more likely to find employment and other opportunities through casual acquaintances than they are through close friends.
Linkedin issued a statement in which it claimed that over the course of the research, it had “behaved consistently with” the user agreement, privacy policy, and member settings that the firm had in place. According to LinkedIn’s privacy policy, the company utilises the personal data of its users for research reasons.
When asked how the firm had addressed the possible long-term repercussions of its trials on the job and economic position of its members, LinkedIn, which is owned by Microsoft, did not explicitly answer the issue. However, the corporation said that the study did not benefit certain customers more than others in a disproportionate manner.
There is a murky past to large internet corporations doing experiments on their customers. Eight years ago, a study that described how Facebook had quietly manipulated what posts appeared in users’ News Feeds in order to analyse the spread of negative and positive emotions on its platform was published. The study described how Facebook had done this in order to analyse the spread of positive and negative emotions on its platform. A reaction was rapidly produced in response to the one-week trial that was performed on 689,003 people.
When individuals joined up for Facebook, the authors of the Facebook study, who included a researcher who worked for the firm and a professor at Cornell, maintained that users had unwittingly given their permission to participate in an experiment involving the manipulation of their emotions.
The other viewpoint was held by critics, with some of them accusing Facebook of having breached people’s privacy while also abusing their emotions and leading them to experience mental discomfort. Others felt that the initiative had used an academic co-author in order to add respectability to potentially dubious business research procedures.
Cornell later stated that its internal ethics board had not been required to review the project because Facebook had independently conducted the study and the professor, who had helped design the research, had not directly engaged in experiments on human subjects. In addition, Cornell stated that the project did not violate any of the university’s policies regarding the treatment of animals.
The programme examines members’ career histories, job titles, and connections to other users, among other types of data. After that, it attempts to determine the chance that a LinkedIn member would extend a friend invite to a proposed new connection and also the likelihood that the suggested new connection will accept the invitation sent by the LinkedIn member.
In order to carry out the trials, LinkedIn modified its algorithm in order to randomly change the proportion of strong to weak relationships that were suggested by the system. According to the findings of the research, the first round of testing, which took place in 2015, included “nearly four million experimental volunteers.” More than 16 million individuals took part in the exams during the second round, which was carried out in 2019.
Professor Aral from the Massachusetts Institute of Technology stated that the more profound significance of the study was that it demonstrated the significance of powerful social networking algorithms. These algorithms are important not only in amplifying problems such as misinformation, but also as fundamental indicators of economic conditions such as employment and unemployment.
Catherine Flick, a senior researcher in computers and social responsibility at De Montfort University in Leicester, England, characterised the study as more of a corporate marketing exercise. She said this in reference to the fact that the research was conducted by the university.
