The impending challenge of marrying Artificial Intelligence with privacy

Artificial Intelligence (AI) is taking the world by storm. In recent years, we have heard about AI-led advances in multiple industries. The automobile sector is using AI to make roads safer through self-driving cars. AI has also been a driving force in the healthcare industry, aiding medical diagnosis and helping in drug creation. It is also contributing to research and training in newsrooms, as evidenced by the recent collaboration between Google and the London School of Economics.

The capabilities of AI not only impact individual life and whole sectors alike, but can also have huge implications for how whole nations approach their quest for progress. For instance, AI can be used to reshape how the food industry is structured around India. Consider a scenario where we might have access to the data of an online vegetable delivery company. We would be able to gauge the demand for different vegetables in different regions. So if a region has high demand for ladyfingers, we could aim to grow the vegetable closer to the market. This would cut down on transport costs, and would ensure that the consumers get fresh(er) product.

Similar insights can have tremendous potential for how developing countries such as ours allocate resources. AI’s algorithms learn from huge data sets that they are fed. The more and diverse data AI processes, the better its insights become. Big Data is essentially a learning tool for AI. It is a mutually beneficial relationship where AI gains learning tools through processing data and big data is able to spill correlations which were previously unthinkable. To sum it up, Big Data includes large-scale recordings of human behavior that are processed by AI to reveal insights.

This relation between AI and Big Data, while seeming simple on the surface, becomes more complicated when we delve deeper, and look at privacy and data protection concerns. The growth of AI has implications for data privacy that require careful attention as we forge a path forward driven by AI’s capabilities.

Anonymized sets of Big Data may not make sense to human eyes, however, feeding enough of these sets to AI will lead to the algorithms identifying the users who generated this data, without ever having been provided their names. Because AI can compare two or more databases, it may not need the name and/or social security/Aadhaar numbers to identify you. Data containing “location stamps” – information with geographical coordinates and time stamps – could be used to easily track the mobility trajectories of how people live and work.

There are two aspects to this issue – one, how can AI enhance privacy for companies that handle large data sets. Second, how can privacy be embedded into development of AI tools and technologies.

With its advanced capabilities, AI can help companies identify and respond in real-time to individuals who are looking at consumer data and prevent inappropriate use or theft of data. In addition, AI can also help develop privacy tools to help the consumer. A good example of the same is the AI, ‘Polisis’ which stands for Privacy Policy Analysis. The application can read privacy policy documents to develop insights such as an executive summary and a flow chart of what kind of data is collected and who it will be sent to. In addition, it also outlines whether or not the consumer can opt out of the collection or sharing of the data.

Data is going to drive the economies of the future, and in a data-driven regime, the idea of privacy takes center stage to protect the interest of consumers and citizens alike.

On the second aspect, privacy by design techniques should be incorporated in AI development. With rising data collection and storage, doctrinal notions around ‘consent’ and ‘privacy notices’ should be considered. The model of ‘clickwrap’ contracts which allows the user to click on “I accept” button without reading a long, verbose and unintelligible privacy terms and conditions needs to be also revisited. Privacy by design techniques can be incorporated at the level of privacy notices but also at each level of information flow till its storage and processing stage. Further notions of transparency, accountability, and fairness must be incorporated.

There can be no strict set of rules or policy guidelines which can bound an algorithm designer, but, best practices following constitutional standards jurisdiction-wise can be developed as a benchmark. A few techniques that could be deployed to enhance privacy of an existing AI tool are differential privacy, homomorphic encryptions and generative adversarial networks. Along with this, another privacy enhancing and data protection measure which should be taken is of certification schemes and privacy seals to help demonstrate the compliance by organizations.

Privacy and AI can not only complement each other but also enhance the overall output while at the same time providing consumers the best user experience while protecting his/her data. The fundamental basis on which privacy by design works is when a risk is identified, it is looked upon as an opportunity, rather than a challenge, to find creative solutions that deliver the ultimate benefits while securing data. AI is proliferating, so it is necessary to embed privacy and appropriate technical and organizational measures for it into the process that lead to positive outcomes. The opportunity therefore, for AI today, is not just solving for corporations and nations, but instead, to do so in a manner that is sustainable in terms of user privacy.

This article is written by Kazim Rizvi, Founding Director, The Dialogue

Go to Source

Leave a Reply

Your email address will not be published.