By Saraswathy Vaidyanathan
From the times when internet and technology were mere additions to our lives, we have now moved towards the world of artificial intelligence, big data, and algorithms becoming an integral extension of our lives. Algorithms have also become pervasive to the extent that, they influence the basic aspects of our lives. They shape the news and information we have access to by way of information retrieval algorithms.
One important type of such retrieval algorithms are search systems. The use of search systems in our daily lives has become indispensable. Any time a person has a question in their mind, they first seek the help of a search engine, most likely Google. The search query can relate to finding a specific website or piece of content (navigational queries) or seeking information about events, topics, or people (informational queries).
While the objective of the search queries is to provide the user with information and knowledge, such information and knowledge might not always be without prejudice. When a query is fed into the search system, a set of relevant links for the query is pulled out from a set of data items such as posts on social media platforms, web links, and newspaper articles. These relevant links are fed into a ranking system. A rank list is then returned to the user who made the query. The supporters of algorithms reason that the algorithmic techniques exclude human bias in the decision making process. Studies highlight that “many users believe that these algorithmically curated channels (as opposed to human editorial curation) are powerful, infallible and thus unbiased” However, the allegations of ‘search bias’ on Google prove otherwise.
Google has been alleged of search bias before the European Commission (EC), the Federal Trade Commission (FTC), and the Competition Commission of India (CCI). This blog post looks at search bias allegations on Google raised before the CCI. For a better understanding of this issue, comparisons have been made at appropriate places, to judgments of the above mentioned authorities.
The Google decision
On February 8th, 2018, the CCI imposed a fine of INR 135.86 crores (approximately 21.17 million dollars) on Google. It was alleged that Google abused its dominant position and created an uneven playing field in terms of specialised searches, by manually manipulating the search results to the advantage of its vertical partners. These partners provide a range of vertical search services such as YouTube (video), Google news (news), Google maps (maps), Google flight (flights). It was alleged that Google tampered with the organic search results by mixing the vertical search results into them to promote its vertical partners.
Further, it was argued that Google was an unavoidable trading partner in the search advertisement market because of its dominance in algorithmic search market. It was also pointed out that Google’s own sites would appear prominently on the search results page irrespective of whether they are the most popular or relevant sites to the search and Google would not place results from any other vertical search sites as prominently as its own vertical search sites in its list of results. Hence, it was claimed that Google abused its dominant position in the relevant market in India and was thus in contravention of Section 4 on the Competition Act of India.
To determine the abuse of dominant position or monopolisation, the CCI followed three steps:
- Delineation of the relevant market;
- Determination of the dominant position of the firm; and
- Assess whether the conduct amounts to abuse of dominant position.
Google was held liable for the same, on the following grounds:
- Displaying ‘universal results’ in fixed positions in the search engine results page (SERP), rather than in the order of relevance;
- Manipulating the search algorithm to favour its own vertical search services which are prominently displayed in the SERP; and
- Imposing unfair conditions in the syndication/intermediation agreements with website publishers.
This legal intervention by CCI in the digital market is seen as the first of its kind and it can be expected that with increasing market for big data, algorithms, such interventions by the CCI will become a regular part of its governance.
It is pertinent to note that this approach has not been the best approach in comparison with how the other jurisdictions have dealt with allegations of search bias. All the three jurisdictions used the abovementioned three steps but arrived at different conclusions. For example, the EC, while finding Google guilty of search bias resorted to analyse user’s clicking behaviour to understand whether competitor’s actually lost out on volume due to Google’s conduct, whereas CCI made no such efforts. Further, the EC also resorted to relying on market studies to understand the user behaviour in EU which showed changes in click patterns vis-à-vis changes in SERP. The judgment by CCI suffers from some major drawbacks, lack of fact finding to support its conclusions; no empirical data to support the findings in the judgement; the amount of consumer harm and competitor harm is not highlighted properly. Additionally, EC made Google recommend the solutions and commitments to remedy the breaches which is a more market oriented approach. On the other hand, CCI’s direction to Google to provide disclaimer in the vertical search box is inadequate to contain consumer harm. These drawbacks, in fact, would make Google’s appeal stronger as the scope for reviewing conclusions is enhanced in the light of no data backing up the findings and further make any hard impact in way of Google’s operations in India.
These differences are also important from the Indian perspective because search bias cannot be solely attributed to Google. On being confronted by the United States (U.S.) Congress, about search bias, Sundar Pichai (CEO of Google) on record, made the argument that “This is working at scale, and we don’t manually intervene on any particular search result”. He further explained the search process as “Any time you type in a keyword, as Google we have gone out and crawled and stored copies of billions of [websites’] ……. and match it against their pages and rank them based on over 200 signals.”
Going by this explanation, clearly the user’s conduct makes a difference. For example, the keywords “Chhota Bheem Indian politics” returns results talking about the politician Rahul Gandhi instead of returning Chhota Bheem cartoon episodes This search result is a clear indication of how user’s perspective/bias finds its way to online search systems. So, perhaps, in future, CCI should take similar steps as EC to arrive at conclusions which would solidify the roots of the working of Indian competition law. However, it is not hard to accept the fact that globally, jurisdictions are grappling with regulating digital space by anti-trust interventions.
The road ahead
Google’s monopoly is a striking feature. The core of all allegations is the abuse of a dominant position. An entity gains a dominant position when there are no alternatives or the general public is not aware of any alternatives. I would argue the latter i.e. the lack of general knowledge of alternatives. An able alternative to Google could be the search engine DuckDuckGo (DDG). “The search engine that doesn’t track you” is the slogan of the search engine. While Google collects search data and links it to personal accounts which eventually is used to personalise search results, DDG does not store search data and also does not personalise search results. So, if more users are aware of DDG, there are high chances of at least more such persons trying to use alternate search engines. This usage will perhaps break Google’s dominant streak and in a way reduce search bias. The pertinent question is whose role is it to create this awareness of alternatives?
Search bias could play an important role in the upcoming 2019 General Election results in India. Studies have shown that political bias could be an extension of search bias. In a study conducted using US Presidential elections as a framework, results show that the top search results could influence user’s opinion and public opinions would be result of just certain perspectives than a wholesome opinion. In the present political climate of India, where fake WhatsApp forwards are setting the tone of elections, fighting political bias would be a real challenge. Hence, at present times, it is all the more necessary that more empirical studies are conducted to understand the phenomenon of search bias. The aspects overlooked by the CCI while arriving at conclusions should form the base for such studies. Based on the studies, steps can be taken to further strengthen the implementation of competition laws in a manner that makes India stand at par with its global counterparts to hold giants like Google accountable to the maximum extent.
The author would be joining CIIPC as a Research Fellow. Views are personal.
M.J. Welch et al, Search result diversity for informational queries, (2011). In: Proceedings of the 20th international conference on World Wide Web. ACM, 237-246.
 Juhi Kulshrestha et al, ‘Search bias quantification: investigating political bias in social media and web search’ (2018) Information Retrieval Journal 417-432
 Solon Barocas and Andrew D. Selbst, ‘Big Data’s Disparate Impact’ (2016) 104 California Law Review 671
 Motahhare Eslami, et al, First i “like” it, then i hide it: Folk theories of social feeds. (2016). In: Proceedings of the 2016 CHI conference on human factors in computing systems ACM, 2371–2382
Also see, A. Springer et al., Dice in the black box: User experiences with an inscrutable algorithm. (2017). In: Proceedings of the AAAI 2017 spring symposium on designing the user experience of machine learning systems. AAAI.
 Competition Commission of India, Cases Nos. 7 and 30 of 2012, Matrimony.com Limited v. Google, 8 February 2018, available at http://www.cci.gov.in/sites/default/files/07%20%26%20%2030%20of%202012.pdf. Para 3
 Id. Para 4
 Id. Para 6
 Id. Para 5
 Id. Para 8
 Id., Para 253
 European Commission, CASE AT.39740, Google Search (Shopping), 27 July 2017, available at http://ec.europa.eu/competition/antitrust/cases/dec_docs/39740/39740_14996_3.pdf, paragraphs 454 to 461
 Rohan Jaitly, ‘Donald Trump and Idiot: Sundar Pichai reveals why a picture of the US Prez comes up when you Google “idiot”’ (Times Now, 12 December, 2018) https://www.timesnownews.com/technology-science/article/donald-trump-and-idiot-sundar-pichai-reveals-how-a-picture-of-the-us-president-comes-up-when-you-google-idiot/329846 accessed 21 February 2019.
 Jason Evangelho, ‘Why You Should Ditch Google Search And Use DuckDuckGo’ (Forbes, 3 October 2018) https://www.forbes.com/sites/jasonevangelho/2018/10/03/when-does-googles-convenience-turn-creepy-let-me-duckduckgo-that-for-you/#7d32129e235e accessed 21 February 2019.
 Juhi Kulshrestha et al, ‘Search bias quantification: investigating political bias in social media and web search’ (2018) Information Retrieval Journal 417-432. Also see, R.Epstein and R.E. Robertson, R. E. The search engine manipulation effect (SEME) and its possible impact on the outcomes of elections (2015). In: Proceedings of the National Academy of Sciences (PNAS), 112(33), E4512–E4521.