This article discusses liability for AI systems de lege lata and de lege ferenda on the basis of European and German law with special consideration of the EU Commission’s regulatory proposal for a directive on liability for artificial intelligence (AI). After a brief introduction of the current state of legislative proposals concerning AI and liability (I.), the characteristic features of AI system and their relationship with the concepts of AI system as well as the requirements for human fault for liability are carved out (II.) and different liability regimes considered (III.), before the Commission’s approach to liability for AI taken in its proposal for a directive is thoroughly analysed in the light of its impact on the gathering evidence and the burden of proof in liability claims (IV.) and the findings are summarised (V.).

This article considers the key provisions that make up this proposed lite framework, notable deviations from prior iterations and what this new age of data protection will likely mean for citizens and businesses alike. In stark contrast to the comprehensive General Data Protection Regulation (GDPR) – still widely considered the ‘gold standard’ in data protection – the draft Digital Personal Data Protection Bill (DPDP Bill) adopts a simpler, and arguably, more simplistic approach to protecting personal data.

This article comparatively reviews three precedents (Worth, Southeast Power and Barrett) concerning failed enterprise resource planning implementations, compares the role of specific key themes in those enterprise resource planning implementation failures and highlights the key lessons learned.