Digital Ethics and the Right to Be Forgotten

Why It Matters

The concept of a "right to be forgotten" emerged from European data protection law, but its philosophical foundations run far deeper than any statute. At its core, this right asserts that individuals should have some control over their digital past, that the permanence of online information fundamentally alters the human experience of time, memory, and personal reinvention. In an era where a single photograph, social media post, or data breach can define a person indefinitely, the philosophical stakes of digital memory are immense.

Kazanskaia (2025) frames the challenge around three interlocking concepts: privacy as a condition for autonomy, ownership as control over one's digital representation, and governance as the institutional structures that mediate between individual rights and collective information needs. The tension between these dimensions is what makes digital ethics so philosophically rich and practically intractable.

The arrival of large language models adds an entirely new dimension to this debate. Kazanskaia (2025) demonstrate that even when knowledge is supposedly "erased" from LLMs through unlearning techniques, step-by-step reasoning attacks can recover the deleted information. This means that the right to be forgotten may be technically unenforceable in AI systems that have already ingested personal data, creating a philosophical crisis for data protection frameworks built on the assumption that deletion is possible.

The Debate

Philosophical Foundations of Digital Forgetting

The right to be forgotten draws on several philosophical traditions. From a Kantian perspective, it protects the dignity of persons by preventing them from being reduced to their worst moments. From a virtue ethics standpoint, it preserves the possibility of moral growth and character development. Alekseeva and Rokotyanskaya (2025) argue that digital ethics requires a new philosophical synthesis because existing moral frameworks were developed for a world where information naturally decayed over time.

The Tension with Collective Memory

Not all forgetting is benign. Societies need historical records to hold powerful actors accountable, to learn from past injustices, and to maintain shared narratives. The philosophical challenge is distinguishing between legitimate claims to personal reinvention and attempts to escape accountability. A politician who wants embarrassing policy positions erased presents a fundamentally different case from a teenager whose youthful indiscretion has become permanently searchable.

Youth, Autonomy, and Digital Permanence

Shouli et al. (2025) highlight the particular vulnerability of young digital citizens whose data is collected before they can meaningfully consent. Children's digital footprints are created by parents, schools, and platforms long before they develop the capacity for autonomous decision-making. This raises profound questions about intergenerational digital justice and the right of future adults to shape their own digital identities.

The AI Unlearning Problem

The technical discovery that LLM knowledge erasure is unreliable creates a philosophical paradox. If an AI system has learned facts about a person from training data, and those facts cannot be reliably deleted, then the right to be forgotten becomes a legal fiction when applied to AI. This may require fundamentally rethinking data protection from a deletion-based model to an access-control model, or it may demand that AI systems never ingest certain categories of personal information in the first place.

Digital Forgetting: A Rights Framework

What To Watch

The intersection of the right to be forgotten with generative AI is the frontier to monitor. As regulators grapple with how GDPR-style deletion rights apply to model weights that encode personal information in distributed, non-retrievable ways, expect new philosophical frameworks that move beyond binary present/absent conceptions of data toward probabilistic and contextual models of digital identity governance.

Why It Matters

The concept of a "right to be forgotten" emerged from European data protection law, but its philosophical foundations run far deeper than any statute. At its core, this right asserts that individuals should have some control over their digital past, that the permanence of online information fundamentally alters the human experience of time, memory, and personal reinvention. In an era where a single photograph, social media post, or data breach can define a person indefinitely, the philosophical stakes of digital memory are immense.

Kazanskaia (2025) frames the challenge around three interlocking concepts: privacy as a condition for autonomy, ownership as control over one's digital representation, and governance as the institutional structures that mediate between individual rights and collective information needs. The tension between these dimensions is what makes digital ethics so philosophically rich and practically intractable.

The arrival of large language models adds an entirely new dimension to this debate. Kazanskaia (2025) demonstrate that even when knowledge is supposedly "erased" from LLMs through unlearning techniques, step-by-step reasoning attacks can recover the deleted information. This means that the right to be forgotten may be technically unenforceable in AI systems that have already ingested personal data, creating a philosophical crisis for data protection frameworks built on the assumption that deletion is possible.

The Debate

Philosophical Foundations of Digital Forgetting

The right to be forgotten draws on several philosophical traditions. From a Kantian perspective, it protects the dignity of persons by preventing them from being reduced to their worst moments. From a virtue ethics standpoint, it preserves the possibility of moral growth and character development. Alekseeva and Rokotyanskaya (2025) argue that digital ethics requires a new philosophical synthesis because existing moral frameworks were developed for a world where information naturally decayed over time.

The Tension with Collective Memory

Not all forgetting is benign. Societies need historical records to hold powerful actors accountable, to learn from past injustices, and to maintain shared narratives. The philosophical challenge is distinguishing between legitimate claims to personal reinvention and attempts to escape accountability. A politician who wants embarrassing policy positions erased presents a fundamentally different case from a teenager whose youthful indiscretion has become permanently searchable.

Youth, Autonomy, and Digital Permanence

Shouli et al. (2025) highlight the particular vulnerability of young digital citizens whose data is collected before they can meaningfully consent. Children's digital footprints are created by parents, schools, and platforms long before they develop the capacity for autonomous decision-making. This raises profound questions about intergenerational digital justice and the right of future adults to shape their own digital identities.

The AI Unlearning Problem

The technical discovery that LLM knowledge erasure is unreliable creates a philosophical paradox. If an AI system has learned facts about a person from training data, and those facts cannot be reliably deleted, then the right to be forgotten becomes a legal fiction when applied to AI. This may require fundamentally rethinking data protection from a deletion-based model to an access-control model, or it may demand that AI systems never ingest certain categories of personal information in the first place.

Digital Forgetting: A Rights Framework

| Individual | Personal reinvention | Erasure of past data | vs. public accountability | | Society | Collective memory | Access to historical record | vs. individual dignity | | Platforms | Data monetization | Retention for services | vs. user autonomy | | AI systems | Training completeness | Learned knowledge persistence | vs. technical erasure limits | | Youth | Future autonomy | Deletion of childhood data | vs. parental/institutional records |

What To Watch

The intersection of the right to be forgotten with generative AI is the frontier to monitor. As regulators grapple with how GDPR-style deletion rights apply to model weights that encode personal information in distributed, non-retrievable ways, expect new philosophical frameworks that move beyond binary present/absent conceptions of data toward probabilistic and contextual models of digital identity governance.