Online connectivity now defines our 'information civilisation' and presents many benefits and risks. The dynamics of these multi-layered risk/benefit relationships are complex but what is common throughout are risks relating to metrics of increasing values, from number of users connected, types of connectivity, time users spend connected, the number of connected devises, and the increase in user data harvesting. The online phenomenon presents an increas- ingly complex risk phenomenon. Fortunately, research confronts many of these risk contexts, so much so there are many growing narratives of both benefits and risks regarding online connectivity. The article focuses on one particular narrative concerning the risks of the connected online phenomenon. For the ease of discussion, we use Sherry Turkle’s 2006 work the "Tethered Self" as the start of the online connectivity and risk narrative. Turkle framed some of the risks of increasing connectivity, under the title of the “always-on culture”. The narrative has grown in recent times with the addition of the internet of things as another medium of connectivity consisting of numerous forms of “always on devices”. The article maintains that the growing popularity and development of artificial intelligence assistants presents another evolutionary sequence of the always-on narrative. Furthermore, the narrative now moves from user con- trolled connectivity, third party connectivity to connectivity mediated through artificial intelligence assistants/agents. The article aims to interrogate and contribute to the risk framing of artificial intelligence assistants by situating the technology in the always-on narrative.