ForumIAS announcing GS Foundation Program for UPSC CSE 2025-26 from 27th May. Click Here for more information.
Source: This post is created based on the article “Laws should keep pace with fast evolving technologies” published in Live Mint on 26th April 2023.
Syllabus Topic: GS Paper 3 – Awareness in the fields of IT
Context: Recent developments in the field of AI are conflicting with Europe’s GDPR.
Recently, discussions on regulating artificial intelligence (AI) have raised concerns about privacy. Many experts believe that AI technologies, like Large Language Models (LLMs), may not meet the strict privacy requirements of Europe’s General Data Protection Regulation (GDPR). Since GDPR is considered a global standard, this could impact AI technology deployment worldwide.
What are the conflicts between GDPR and Large Language Models?
GDPR emphasizes obtaining individual consent before collecting and processing personal data. LLMs, however, training data is gathered from the web, which potentially includes personal information without consent. GDPR allows data collection without consent if there is a “legitimate interest,” but it is difficult to argue that LLMs are a “necessity” in this context.
Moreover, GDPR principles of data minimization and retention restriction allows collection of that much personal data which is required to achieve a specific purpose. The collected data should be removed after that. This principal conflicts with LLMs’ reliance on vast and indefinite data availability for continuous model refinement. only so.
Lastly, users could also share their personal info with Conversational AI solutions. It becomes a part of the learning cycle and becomes part of the AI model. It also risks gathering personally identifiable information, potentially violating GDPR.
What should be the course of action?
Recently, many European country regulators are investigating the matter of privacy involved in the OpenAI’s use.
However, considering these issues, some suggest re-examining our present legal frameworks to better align with new technologies like LLMs and generative AI.
As past technological advancements have also prompted legal changes, it is crucial to strike a balance between innovation and privacy without forcing new technologies to comply with outdated frameworks.
