Your Digital Shadow

In our digital lives and society, everything is represented in data. This gives us two versions of the world: the physical one and its digital reflection, like a skewed mirror image, similar, but not completely the same.

More and more, it’s not the physical you but the data version of you that others interact with—whether it’s a bank checking your credit, an algorithm showing you certain search results, or a potential employer “stalking” you online. Our interactions with the world are increasingly shifting from the physical realm to the digital one, where we have less control over what happens or how we’re perceived.

This raises several ethical and philosophical concerns. What happens when we treat people as collections of data? What happens when our digital profiles determine whether we’re accepted or rejected in society? The datasets that are available and that describe people can fundamentally change how we are perceived.

The idea of describing our digital shadow originates from privacy advocates. A shadow has connotations analogous to our digital selves: It’s something that follows us, that hides things, and that isn’t necessarily a good, and certainly not a complete, representation of who we really are.

The digital shadow is a virtual representation of you. It is constantly updated through your actions and interactions, as a result of both active and passive data analysis, collected from both what we consciously share and that which we can’t control ourselves.

This digital representation isn’t final and fixed, rather it evolves, changes, and follows you continuously. The idea is a variation of what’s known as a digital twin.

When decisions are made that directly or indirectly affect us, they are often based on our digital representations, whether it’s a person or an algorithm such as an AI system making those decisions.

This isn’t entirely a bad thing. Digitalisation has, for example, made tasks such as applying for loans or filing taxes much easier, as it automatically gathers necessary information from various sources, so we don’t need to frantically search through old documents, receipts, and accounts.

These are important advancements, and as we have become accustomed to them, we can’t simply go back to the old ways. However, we must be aware of and work to minimise the potential downsides.

Challenges include the possibility of data being incomplete or of poor quality. Also, data is self-referential, meaning it can only point to other data; it is disconnected from the real world. The data about you is stored in various places and to a certain extent has “a life of its own”, beyond your control. If decisions are made automatically and it’s not transparent on what basis the decision was made, it becomes difficult to verify their accuracy.

Moreover, the treatment will never be completely neutral, it will be shaped, consciously or unconsciously, by the values, prejudices, and agendas of those who have developed a specific service and those who are running it.

If it involves clear-cut information, such as income and debt on a tax return, it’s generally straightforward, transparent, and correctable. In cases that involve more assessments and interpretations, which may be run by private companies, and where you have less insight or control over what’s happening, this may not be so.

If these errors and shortcomings lead to consequences such as denial of financial aid or exclusion from an educational program, they impact your real-world situation in a negative way. We do have some protection against this, thanks to European laws such as the GDPR, which states that everyone has the right not to be subject to a decision based solely on automated processing. However, there are many grey areas surrounding these laws, who furthermore struggle to keep up with technological development and potential uses of data. We will look at this in more detail soon.