by Carolina Paulesu
It is nowadays a fact that consumers interact with Artificial Intelligence (‘AI’) on a daily basis
by Carolina Paulesu
The interaction with it may take different forms: depending on whether it is used by businesses or by consumers themselves and depending on the use that is made of it, AI can result in potential risks but also bring about benefits for the parties involved.
In the context of online transactions – concerning the purchase of either goods or services – AI proves to be a useful tool both for businesses and consumers.
The ability to process Big Data allows businesses to extract information and identify people’s consumption patterns.
By doing so, they can adapt their conduct so to improve the likelihood of consumers adopting a certain behaviour, but also modify their own business strategies and improve them.
This is the case, for instance, of targeted advertising, based on consumers’ online behaviour, which consists in addressing advertisements relating to a product or a service to a specific audience who is predicted to be interested, thanks to the processing of data (consciously, or not) disclosed by consumers online.
This practice brings advantages for consumers as well: they are offered products and services which are relevant, rather than not, and helps them navigate and discern the overwhelming extent of information available online.
Hence, while certain applications of AI may benefit consumers, it cannot be ignored that often – far from limiting themselves to mere nudging – businesses exploit the information obtained through data, by leveraging consumers’ vulnerabilities and cognitive biases.
Targeted advertising itself can prove to be harmful: think of predatory loans offered to consumers who find themselves in distress from a financial point of view.
Another practice possibly damaging for consumers is price discrimination.
The latter does not raise particular concerns when it is limited to the physiological maximization of profits, but can become harmful when the differentiation of the price is based on personal characteristics or vulnerabilities of the consumer concerned, or else exploits in a non-transparent manner a given temporary need and the related willingness of a consumer to pay a higher price, all other conditions being equal(e.g., a person who is in need of specific medical treatments and who is, necessarily, less careful about the price offered or that searches several times in a short period of time for information on air-fares for a given travel destination for her holidays).
In this connection, the organic information asymmetry between businesses and consumers may be exacerbated by AI, and lead consumers into taking commercial decisions they would not have taken otherwise, had they been aware of the process behind and of the discrimination that have been subjected to.
From this point of view, the normative area at which to look in order to find remedies to the potential distortions generated by such uses of AI is primarily represented by the system of protection of personal data.
Yet another source of regulation may be provided by the rules that in general protect consumers from unfair contractual clauses, including in the event of lack of transparency in the pre-contractual phase, or from those transactional forms that may qualify as unfair commercial practices.
AI applications described above are characterised by being used by businesses. However, in the last few years, multidisciplinary research groups have developed AI tools that may be used by consumers or by consumer organisations. The European University Institute, for instance, created the tool CLAUDETTE which – powered by machine learning – allows the analysis of B2C contracts and privacy policies and detects potentially unfair terms.
All in all, conceiving Artificial Intelligence exclusively as a source of risks for consumers does not seem to be justified. Although its deployment by businesses gives rise to concerns, the potential advantages originating from it when used by consumers cannot be neglected and certainly deserve further investigation.