Oracle: AI agents decide and act. The question of liability still remains

Oracle says it is building a range of AI agents into its cloud-based enterprise applications and claims they can make and execute decisions autonomously within business processes. However, given the unresolved questions about data integration and liability, analysts advise caution.

Launched this week in London, Fusion Agentic Applications will be integrated into the Oracle Fusion Cloud Applications suite, which covers finance, ERP, human resources, payroll and supply chain management. Oracle argues that it has a structural advantage here: the data needed to train and operate these agents already resides in its enterprise applications.

“Applications that can think, decide and act to pursue defined business goals,” is how Steve Miranda, chief application developer at Big Red, described the change, a move away from process-oriented software towards results-oriented automation.

Oracle, for example, promises design-to-source…