Beyond the Algorithm of AI - A J Ecosystem Paper

Artificial intelligence is supported by an interconnected commercial ecosystem involving hardware, data centres, electricity, datasets, foundation models, cloud platforms, applications, intellectual property, finance and end users. Different entities may own the hardware, operate the infrastructure, collect or license data, develop models, supply applications and contract with customers. These activities may take place in different jurisdictions and may be organised through independent suppliers or related-party arrangements.

This paper follows the commercial chain and examines the GST, Customs, direct-tax, international-tax, transfer-pricing, FEMA, data-protection, intellectual-property, cybersecurity, competition, consumer-protection and sectoral rules that may apply at each stage. It also considers India's AI policy architecture, including the IndiaAI Mission, public compute infrastructure, industrial policy and the developing governance framework.

The legal analysis begins with identifying what is actually supplied, which rights are granted, who performs the relevant functions and where the parties and activities are located. An AI offering may combine imported hardware, cloud computing, model access, software licensing, data processing, implementation and human support. These elements may require separate analysis even when the customer receives a single integrated solution.

The paper distinguishes existing law from policy guidance, proposals and unresolved questions. This is particularly important in relation to AI governance, copyright and training data, the phased commencement of India's digital personal-data protection framework, GST treatment of cloud and data-hosting services, and the characterisation of cross-border AI payments. These issues should be analysed under the law applicable to the relevant transaction and period, rather than through generalisations about AI as a technology.