Is it plausible to develop targeted online marketing that does not impinge on the consumer’s privacy? Of course, the dilemma is that in order to show putative customers advertisements that are likely to hit the mark rather than random enticements, the advertiser needs to know something about the putative customer’s interests, tastes, personality, and other factors. Many see the acquisition of such information as worrying and there are now many places where gathering personal and private data would be in breach of local laws.
Writing in the International Journal of Social Computing and Cyber-Physical Systems, a team from India discusses the very possibility of privacy-preserving, but nevertheless, targeted advertising. Ainish Dave, Hardik Gajera, and Manik Lal Das of DA-IICT, in Gandhinagar, India, explain that targeted advertising is very attractive to sellers, because its success is likely to be greater than non-targeted, almost advertising. They too worry about a consumer’s private data getting into the hands of third parties. As such, they are developing a new model for targeting that does not compromise one’s privacy.
In the new approach, only keywords are extracted from the user’s browser history and these are encrypted homomorphically before being stored within the system to allow a targeted advertisement to be selected for that user without their personal data being acquired and stored on the advertising company’s servers and without anyone actually knowing what keywords were used to pull in the advertisement. Tests so far show that it is efficient and practical even when compared to other available advertising technology. “The security analysis of the proposed scheme shows that the scheme is secure with the hardness assumption of approximate-GCD problem, which is an intractable problem,” the team reports.
Dave, A., Gajera, H. and Das, M.L. (2019) ‘Privacy-preserving targeted online advertising‘, Int. J. Social Computing and Cyber-Physical Systems, Vol. 2, No. 2, pp.132-145.