As artificial intelligence (AI) continues to revolutionize industries and reshape society, its implications for privacy are becoming increasingly urgent. Among the most pressing concerns is the challenge AI presents to the Right to Be Forgotten (RTBF).

Originally designed to protect individuals by allowing them to request the deletion of their personal data, the RTBF now faces significant obstacles when applied to AI systems.

Bashirat Atata, a prominent data privacy expert and UC Berkeley Law graduate, recently addressed these issues in her research paper, AI and the Right to Be Forgotten. In her analysis, Atata examines how current privacy laws struggle to keep up with the complexities of AI and offers potential solutions for protecting individual rights in this rapidly evolving landscape.

One of the key challenges identified by Atata is the way AI systems process and use data. Unlike traditional data systems where personal information can be isolated and deleted upon request, AI models integrate data in ways that make it nearly impossible to extract specific information without disrupting the overall functionality of the model.

In AI, personal data is often used to train machine learning algorithms, becoming deeply embedded within the system. As a result, complying with RTBF requests is not as straightforward as it is in traditional systems – isolating and removing personal data could impair the effectiveness of AI models.

AI systems process personal data through various stages, including data collection, storage, and analysis. During data collection, AI systems gather information from multiple sources, such as social media, mobile apps, and IoT devices. This data is then stored in databases, often in a decentralized manner across multiple servers or cloud platforms. The storage of personal data must comply with data protection regulations, ensuring that data is securely stored and accessible only to authorized personnel (Pasquale, 2015). During the analysis phase, AI algorithms process the stored data to extract insights, identify patterns, and make predictions. This process often involves data transformation techniques, such as anonymization or pseudonymization, to protect the identities of individuals. However, even with these techniques, re-identification risks remain, particularly when data from multiple sources is combined. Furthermore, the replication of data across different systems and platforms complicates the enforcement of the Right to Be Forgotten, as it becomes challenging to track and delete all instances of the data (Gasser & Almeida, 2017).

Atata notes that both the General Data Protection Regulation (GDPR) in Europe and the California Consumer Privacy Act (CCPA) attempt to provide mechanisms for individuals to request the deletion of personal data. However, neither of these laws was developed with AI in mind, and as a result, they encounter significant challenges when applied to AI-driven systems. While the GDPR is more comprehensive in its scope, it faces technical limitations in handling AI’s data processing methods. The CCPA, though more flexible, includes broader exceptions that may prevent the deletion of certain data, even when requested. Both frameworks, Atata argues, need to evolve to address the decentralized, global nature of AI data processing.

Recently, the European Union enacted the AI Act, the first major law with AI-specific regulation, signaling a step forward in addressing these challenges. The EU AI Act introduces a risk-based regulatory framework, classifying AI systems according to their potential risk to individuals and society. High-risk systems, such as those used in critical sectors like healthcare and law enforcement, are subject to stricter requirements, including the need for transparency, accountability, and stringent data management practices. This law acknowledges the unique challenges posed by AI and takes steps to mitigate risks while promoting innovation.

In her paper, Atata proposes several potential solutions to bridge the gap between AI innovation and privacy protection. She advocates for the creation of more AI-specific regulations like the EU AI Act that take into account the unique technical challenges posed by data management in AI systems. For example, emerging technologies such as differential privacy and selective data deletion could offer a way to protect individual privacy without undermining the functionality of AI models. Atata emphasizes that these technical solutions should be complemented by legal reforms that provide clearer guidelines for AI developers and users on how to handle personal data responsibly.

Moreover, Atata stresses the importance of international cooperation in addressing AI’s impact on data privacy. AI operates across borders, and regulatory frameworks must reflect this reality. Global efforts to harmonize data protection laws and develop shared standards for AI data management will be essential in safeguarding privacy in a world increasingly driven by AI.

Looking to the future, Atata envisions a balanced approach to AI and data privacy, where innovation can flourish without compromising individual rights. She believes that by refining existing legal frameworks and embracing new technical solutions, society can ensure that privacy remains a priority in the age of AI. Ultimately, her research underscores the need for proactive and adaptive policies that can keep pace with the rapid developments in AI technology.

Bashirat Atata’s work highlights the growing tension between AI advancement and privacy rights, offering a clear call to action for regulators, technologists, and policymakers. As AI continues to shape the future, reimagining the Right to Be Forgotten is critical to ensuring that personal data remains protected in a data-driven world.