Control Of AI: The Collingridge Dilemma
- IJLLR Journal
- 7 minutes ago
- 2 min read
Padma Iyer, Government Law College, Mumbai
ABSTRACT
The Collingridge Dilemma is a term used for the erratic behaviour posed by Artificial Intelligence (“AI”). A newly launched AI tool is generally difficult to control at the start due to the unpredictability of its consequences, but when such impacts come to light, it proves to be harder to mitigate. The reason for this - as elaborated below - is the laissez-faire approach adopted by: first, the developers right after the launch of an AI tool or chatbot; and second, the government, in its insufficient methods of regulatory oversight. Generative AI is only run by a small number of companies, inter alia including Google, OpenAI and Microsoft due to their extensive resources and technological knowhow. However, the issues with AI such as bias has left even the large tech companies helpless, leading to many chat-boxes namely ‘Tay’ and ‘Grok’ being shut down not too long after their launch. In a bid to tackle the growing concerns around AI, this paper firstly explores ways to reduce - if not mitigate - the dilemma posed by AI, such as Bias or Intellectual Property Infringement. The measures elaborated here are: the Precautionary Principle, Responsible Research and Innovation, and Adaptive Governance. Secondly, it is imperative to note that strict rules related to regulatory oversight cannot be implemented immediately. As AI evolves, it is inevitable to see a parallel rise in the disputes related to its usage and development. Therefore, an effective dispute resolution system must be integrated for timely resolution of disputes. This paper delves into the advantages of arbitration for determining such disputes, as opposed to national courts.
Keywords: Artificial Intelligence, David Collingridge, Dilemma of Control, Technology, Arbitration.
