I had a small though experiment about how to model a system influenced heavily by social media that tracks users and builds extensive profiles of them.
On first level, such a system (let's call it A) is equivalent to a imaginary loaded die with a control that
lets one control what comes out when rolling it. A system (let's call it B) without social media is more like a fair die. On a long run, the net outcome of system A will be heavy with induced biases and that of system B will be fair but could be with slight biases due to natural randomness. If the system A is controlled by private interests rather than societal welfare, it would lead to unhealthy societal setup. Hence, the system A is inherently flawed and is also dangerous to overall well-being. System B doesn't have this flaw inherently.
On first level, such a system (let's call it A) is equivalent to a imaginary loaded die with a control that lets one control what comes out when rolling it. A system (let's call it B) without social media is more like a fair die. On a long run, the net outcome of system A will be heavy with induced biases and that of system B will be fair but could be with slight biases due to natural randomness. If the system A is controlled by private interests rather than societal welfare, it would lead to unhealthy societal setup. Hence, the system A is inherently flawed and is also dangerous to overall well-being. System B doesn't have this flaw inherently.
... PS: purely thoughts and so abstract.