Artificial Images and Imaginary of Care. In

The so-called algorithmic turn has given rise to sophisticated image generation systems. Coupled with the growing interest in visual culture, such systems appear to be spreading exponentially. Such a scenario is intertwined with the risks and dangers that such systems can open up for individuals and communities. These include not only privacy and control over data but also biases—those prejudices that transpire “unconsciously” from artificial images and derive from the way algorithms operate by processing iconographic reference datasets, a distorted mirror of society. In particular, by experimenting with these generative systems in relation to professional categories dedicated to care, such as, for example, physicians, psychologists, nurses, childcare and caregivers of the elderly, and people with disabilities, recurring aspects emerge that pertain to the visual imaginary of care. This imagery can be performative, spreading an opaque and seemingly realistic ideal-type of caregiver or physician. The danger is to corroborate certain biases by affecting individual and collective preferences in, for example, the choice of hiring caregivers, physicians, etc. From an ethical standpoint, an AI ethics rooted in the principles of truth and the notion of limit—together with ethics-by-design frameworks and compliance with normative standards—may, at best, mitigate existing biases, though it may also, at times, generate counter-biases. A possible solution, at the moment, may be intertwined with public awareness strategies so that certain biases are first eradicated in society and, over time, also diminished in algorithmic results. Awareness and co-responsibility in the use of such systems pertain to all those involved, from creators to governments to users. This is essential to counteract biases and stereotypes and to foster more critical engagement with how the imagery of certain professional roles is shaped in terms of ethnicity, age, and gender.