The digital transformation of the enterprise is having a deep impact on ICT technology. AI and the Internet-of-Things (IoT) have been “eating corporate ICT from within”, disrupting and replacing conventional software architectures. Traditional software applications are being replaced with a Smart Enterprise IT ecosystem where AI models play an increasingly important role.

The composition of AI-enabled activities (i.e., acquisition, preparation, pre-processing, and analytics) is an orchestrated process, whose configuration requires skills and experience. Eventually it will be managed and tuned by AI entities working in team with humans. Delivering “AI that manages AI” is a key objective of EBTIC Smart Enterprise area for phase 3. Some major technical challenges must be solved to reach this objective. A key challenge is ensuring AI models’ robustness in adversarial environments, where the dysfunctional behaviour of some actors could keep the Smart Enterprise operating sub-optimally.

In EBTIC’s phase 3, the Smart Enterprise theme will rely on Blockchain’s distributed ledger infrastructure, together with light-weight encryption security techniques, to establish consensus on shared information which is essential for enterprise decision making and process optimization, delivering the desired non-functional properties of AI model computation in terms of privacy, data integrity and protection. The Smart Enterprise theme will also address ethics and legal concerns that may depend on the cultural and regulatory environment where AI is deployed. The Smart Enterprise theme will accompany EBTIC’s stakeholders’ transition to secure, explainable and ethically answerable operational AI, supporting them in the change of data management strategies underpinning this transition.