Dr Fazal Ali
The best lie is that a memory always feels so true. LLMs are trained on huge data scrapes that can sometimes “memorise” personal details. To comply with privacy provisions, developers use machine unlearning to make AI models forget particular data points. But this forced forgetting and unlearning is far from the complexity of human forgetfulness that allows memories to be revived and reworked.
Memory is an act of creation. The audio tape memories of women trafficked as indentured labourers to Golconda, the diaries of enslaved women like Mary Prince, which chronicle the memories of her countless floggings, and the legal records of King’s Bench in London that fix the memories of Pictons’ torture of Luisa Calderón in Trinidad serve to close the gap between historical memory and narrative truth.
Human memory is never a complete duplicate of the original experience. Memory is always a process of destruction and creation. Forgetting, repressing and adding. Some of our best memories may never have happened. In fact, they may have happened to someone else. Many of our impulses and enthusiasms, which may seem entirely our own, have arisen from the prompts of others.
Identical events are experienced differently by each person. And events are differently reexperienced whenever they are recollected. There seems to be no mechanism of the mind for ensuring the truth, or at least the veridical character of our recollections. Memory systems have fallacies, frailties, imperfections, and extraordinary flexibility and creativity.
Creativity requires forgetting. Painters, physicists, and poets produce studies, equations and stanzas as if they were completely new, out of the blue, and compounded by an authentic sense of forgetfulness. We later stumble upon many of our thoughts in diaries and discover how they have been forgotten for years. Often, these moments allow old ideas to be revived and reworked. All of us have experienced such forgettings.
In this sense, we can say that creativity may require forgetting so that our memories and ideas can be reborn and reframed in new contexts and perspectives. False memories of fictitious events are also firmly embedded in our minds as deepfake digital clones, and memes that transport misinformation and disinformation circulate continuously on reels and posts without destinations or border control.
Above these false memories is the fact that human memory is always a simulacrum, a copy with no original.
The existential risk is not about the rise of the machine but about the fall of man through overreliance on AI in decision-making.
This overreliance will erode our ability to make ethical decisions. Human free will is challenged not by predeterminism but by algorithms that shape our choices. Moral attrition in the Age of AI, the gradual erosion of our capacity for independent ethical choices, is a feature of post-modern life as we increasingly delegate authority, decision-making, judgement, and responsibility to AI assemblages. The illusion of autonomy is real.
As AI models become more persuasive, epistemic justice and algorithmic fairness become more elusive. Epistemic justice is about ensuring that the knowledge and experiences of marginalised groups are recognised and valued. Algorithmic fairness, therefore, remains a steep challenge because machine learning systems are fundamentally extractive, historically biased, and mathematically reductionist.
Data inherits the historical memory and the pre-judgements of the geographies from which they are extracted.
AI models are trained on vast datasets that mirror inherited inequalities and intergenerational immobility. Models can learn, amplify, and automate these troubles.
In his “Magnifica Humanitas”, Pope Leo XIV underscores the possibility of “new digital slaveries” and calls for the technology to be morally “disarmed”. His encyclical addresses the unprecedented societal shift brought on by algorithms, data, and LLMs, drawing parallels to the Church’s response to the Industrial Revolution.
The Pontif emphasised that technological progress must serve humanity and never undermine human dignity. He warned that unregulated AI could increase unemployment, accelerate autonomous warfare, and enable invasive digital surveillance.
The Pope is adroitly pointing to, and pointing out, the problem of “averages”.
Algorithmic fairness often relies on statistical parity. Because AI prioritises the statistical majority, it often flattens nuanced, localised human experiences, erasing minority viewpoints in favour of dominant ones, thereby violating the principles of epistemic justice, which can eclipse those of natural justice.
Advanced deep neural networks operate as black boxes, making it practically impossible to trace exactly how an AI reached a specific conclusion, which makes it hard to identify and correct algorithmic unfairness.
Fairness is not a universal, objective mathematical formula. It is a fluid metaphysical concept. Different mathematical definitions of fairness often contradict one another, making it impossible to satisfy all definitions simultaneously.
AI development is heavily concentrated in the hands of a minority. This centralises the “epistemic standard”, meaning the knowledge and values embedded in AI primarily reflect the priorities of privileged geographies.
Dr Fazal Ali completed his Master's in Philosophy at the University of the West Indies. He was a Commonwealth Scholar who attended the University of Cambridge, Hughes Hall; the Provost of the University of Trinidad and Tobago; the acting President of UTT; and the Chairman of the Teaching Service Commission. He is the President of NIHERST and an external services consultant with the IDB.
