Individual Addressability Comes Full Circle with Mobile AI: From People to Data to Agents

Abstract

Through the lens of agents, this article situates mobile artificial intelligence (mobile AI) in relation to earlier and ongoing developments with mobile media and communication, including mobile telephony and mobile broadband. With mobile telephony, individuals are directly addressable to one another (“individual addressability”), whereas with mobile broadband, data are addressable to individuals anytime, anywhere (“data addressability”). With mobile AI agents, data are not just addressable to users, but individually addressable to them and the situation at hand (“individual addressability of data”). As social actors that use machine learning to perform tasks requiring humanlike intelligence, mobile AI agents operate at the intersection of addressability among individuals and data, uniquely supporting the “individual addressability of data” and mobile broadband services. As we explain, this development has implications for how individuals connect with people, data, and agents as social actors.

Is “mobile AI” a major development in mobile media and communication (MMC), or just a shiny object that attracts our attention? As we discuss, mobile-mediated artificial intelligence (mobile AI; Goggin, 2025) is in a liminal stage of diffusion, making it risky to forecast exactly how it will continue to be taken up and change the social ecology (Ling, 2023). However, what we can do at this liminal stage is compare how mobile AI structures access to communication and data with the ways other major advancements in MMC make people and data “addressable” to the user, or accessible on a one-to-one level. To that end, this article lays conceptual groundwork by interrogating how and why mobile AI agents can alter the landscape of addressability among people and data in ways that can be compared to, and differ from, advancements in mobile telephony and mobile broadband. Through the lens of agents, which serve as both interface and social actors, we situate mobile AI in relation to these earlier and ongoing developments with MMC.

First, some definitions are in order. A mobile AI agent is an autonomous software entity that provides anytime-anywhere access to machine-learning systems capable of performing tasks requiring humanlike intelligence. This understanding of mobile AI agents combines the definition of AI tools as computational systems that rely on learning to make rational choices and actions (Russel & Norvig, 2021), with that of agents as autonomous software entities that choose actions to help users fulfill their goals (Maes, 1993). Crucially, we also incorporate anytime-anywhere access as a boundary condition for an AI agent to be mobile, which is important because it means its use can be weaved into the flows of the user's everyday life movements and moments.

It is important to note that this analysis may be treated as a trajectory of change in how MMC supports addressability only by recognizing that these developments are layered and emergent, rather than sudden or linear. T-9 predictive texting employed the foundational techniques of machine learning, so one could argue that mobile AI has been around since at least the mid-1990s (Agarwal & Arora, 2007). In the years since, algorithms that personalize data streams have also emerged as early forms of mobile AI. Although it has been emerging for some time, early applications of AI in mobile media were limited, and only in recent years have we started to see an explosion of mobile AI applications and imaginaries from industry, media, and academia (including this special issue). Therefore, it is the building momentum toward mobile AI that marks the moment, not the introduction of mobile AI as something brand new.

As we present this overview, it is also important to bear in mind that there is a great deal of unevenness across societies and populations regarding the adoption and use of MMC, especially mobile AI at this liminal stage. As an overview, this essay offers a high-level perspective on mobile AI that recognizes addressability as a concept that threads through previous MMC developments; it also provides traction for considering how mobile AI agents might extend this concept. As we discuss, this vantage helps lay a foundation for scholarship on the social uses and consequences of mobile AI agents, while expanding the utility of addressability as a concept that explains how and why MMC is meaningful for social life as it continues to evolve with mobile AI.

Mobile Telephony Supports Addressability among Individuals

In the domain of telecommunications networks, “addressability” means that an individual device makes itself identifiable to other individual devices operating on the same network (Farley, 2005). Having identifiers on devices allows them to establish networked connections on a one-to-one basis, which represents a structural shift from the one-to-many broadcast model. We revisit this technical understanding of addressability not to imply technological determinism, but to get at the roots of its meaning, as it has also been taken up to explain the social uses and implications of early mobile telephony.

Early mobile telephony supported voice calling and messaging on a point-to-point basis, and individuals were able to make themselves addressable to other individuals anytime and anywhere they had access to a connected device. Ling and Donner (2009) argue that it is this anytime-anywhere individual addressability that makes mobile telephony not just technically but also socially meaningful. To be sure, anytime-anywhere individual addressability through calling and messaging has brought about changes in the social ecology. To elaborate on one prominent example, when people can reach one another between and beyond places, not just within them, social coordination no longer requires planning around time and location, meaning individuals can “microcoordinate” arrangements iteratively in real time (Ling & Yttri, 2002). Microcoordination is just one example of how anytime-anywhere individual addressability has implications for the social ecology. Individual addressability through mobile telephony has also reconfigured how people engage in personal relationships, family roles, business and commerce, civics and politics, and many other aspects of everyday life (see Ling, 2012).

Mobile Broadband Makes Data Addressable to Individuals

Whereas mobile telephony allowed people to always be addressable to each other, mobile broadband expanded the scope to also make data addressable to people anytime, anywhere. Mobile broadband refers to packet-switched internet protocol connectivity over cellular networks, which allows for access to the internet, as well as cloud- and platform-based services (International Telecommunication Union, n.d.). With early precursors on second-generation cellular networks, mobile broadband became widespread with the uptake of the smartphone and third-generation mobile infrastructure, giving users access to the internet, apps, platforms, and enhanced navigation features anytime, anywhere. However, as we elaborate below, there is an important distinction between an individual having access to data services and data being put into action for an individual. Without the assistance of AI, one still has to navigate and filter their way through mobile broadband environments to make them individually useful. Despite this, having anytime-anywhere access to mobile broadband has been changing the social ecology.

Going back to the example of social coordination, having constant access to mobile broadband restructures the need for maps and directions during microcoordination. Rather than traditional planning, people can use the internet and apps to access information to figure out, in real time, where to go and how to get there. Even before algorithms played a role in dynamically mapping routes and paths, mobile broadband made it possible to access travel and mobility information on the go, restructuring how people travel and get around. With anytime-anywhere addressability of data, many preplanning activities, such as specifying routes, locations, and plans, could be carried out “on the fly.”

Mobile AI Makes Data Individually Addressable

Whereas mobile broadband provides access to data and services, mobile AI takes on the filtering process to make data actionable for individuals anytime, anywhere. Mobile AI harnesses machine learning to perform tasks that normally require human intelligence, which equips agents to filter through data and make it actionable in situations as they unfold. Beyond autonomous calculations (e.g., calculating the shortest distance between locations), AI agents make learned choices for the user (e.g., dynamic routing based on traffic predictions). Increasingly around the world, agents such as Siri, Google Assistant, Manus AI, and Mahindra's Chat Assistant serve as interfaces to AI. Agents are characteristically adaptive (Maes, 1993), and those operating on mobile media will increasingly be able to make data addressable to individuals as they move through environments and situations in daily life (Goggin, 2025). They will increasingly be able to learn the patterns and preferences of their users, detect context, and exert agency to make data actionable to the individual and their situation, including data from the internet, clouds, platforms, the user, and their environment. Without AI, filtering is an intermediary between individuals and actionable data. Although agents become the intermediary, they may perform this role in the background so that it is still experienced as the user having direct access to actionable data. To the extent an agent has the autonomy to make intelligent choices (not just calculations) for the user and act on their behalf, it makes data more individually addressable to them. The agent might be a go-between, but when operating rapidly and behind the scenes, data can be experienced as individually addressable rather than generally addressable, as with mobile broadband.

Individual Addressability among Social Actors

As mobile AI agents make data individually addressable, they also perform the role of communicative actors that interact with people and each other, reconfiguring the “social” contexts of MMC. People commonly use social heuristics when engaging with technologies that have humanlike qualities, such as facial features or a voice (Nass & Moon, 2000). Agents can trigger social heuristics by communicating like humans and with humans, while exhibiting humanlike purpose and intentionality through their agency (Campbell et al., 2025). In fact, several theoretical traditions treat agents as social actors with agency, including the computers are social actors framework (Nass & Moon, 2000), actor network theory (Latour, 2005), and agential realism (Barad, 2007). In addition to making data individually addressable, mobile AI agents may also alter the landscape of individual addressability by adding another layer of communication and social interaction, beyond the dyadic connections characteristic of (traditional) individual addressability (Ling & Donner, 2009). When an agent (inter)acts as an interlocutor, it adds another node to the social structure.

On the one hand, it might seem to make addressability among people less direct when agents are mediating between them. If we think of classic models of communication (e.g., Shannon & Weaver, 1949), agents may have the potential to add “noise” to the communication process by serving as a go-between. In these situations, their involvement might make interactions between people like the game “telephone,” with the agent in the middle adding noise (e.g., suggestions and information) to the flows of communication among individuals. On the other hand, agents may lessen the noise in communication between people. As they become more able to work with each other and across platforms, agents will increasingly be able to coordinate group logistics. Returning to the example of microcoordination (Ling & Yttri, 2002), the iterative back-and-forth of making arrangements in real time can be manageable with two or three people, but with larger groups and networks it can become overwhelming (Ling & Lai, 2016). Existing AI-mediated systems, such as ride-hailing and delivery platforms, demonstrate movement toward agents managing situational complexity on behalf of users. Furthermore, platforms such as Apple, Google, and Open AI are promising next-generation assistants that act on behalf of users to coordinate plans with other people and agents (e.g., Apple Newsroom, 2025). Although we cannot forecast adoption and usage, we can expect these assistants to become increasingly available to manage micrcoordination among groups. Agents accessed through mobile media may be able to detect (e.g., through sensing and data inputs) contextual aspects of the individual (e.g., their psychological state), the physical environment (e.g., spatial and social surroundings), and the media conditions (e.g., affordances and constraints) (Schnauber-Stockmann et al., 2025). In this scenario, much of the communication about situational context (e.g., “Where are you?”; Arminen, 2006) might be moved behind the scenes, expanding the possibilities for real-time microcoordination to occur beyond dyads, among larger groups and networks. Agent-to-agent communication presents a more complex model of individual addressability, with logistics being arranged through software entities that interact with each other behind the scenes, as they also interact in the foreground with individuals and maybe even each other. In this expanded model of individual addressability, microcoordination is not just extended beyond dyads, but also potentially offloaded.

At the Intersections of Individual and Data Addressabilities

Considering agents connect users with both people and data, it is important to consider mobile AI at the intersection of social and data addressability. As discussed above, agents will increasingly be able to filter through digital ecosystems to make data actionable to individuals in the situation. When these situations involve people interacting with other people, mobile AI agents are using the individual addressability of data to support individual addressability among people. We can see this already with some AI calendar agents, such as Motion's (n.d.), which can monitor multiple calendars and account for travel time and location changes to rearrange meetings automatically. This integrated model of individual and data addressability could have distinctive implications for how people connect and interact as they go about their daily lives. Offloading the back-and-forth of microcoordination while scaling it up to groups is just one example where we might expect to see changes in routines and practices, if mobile AI agents are taken up in this way.

We want to stress that this is a big “if”: it is important to remember that these agents, and mobile AI more broadly, are presently in a liminal stage, and we do not know exactly how ongoing developments might be taken up and lead to social change. Our focus thus far has been to interrogate the possibilities of mobile AI agents for addressability, rather than to forecast their actual uses and social consequences. We conclude with an approach for reconciling the potential gaps between what mobile AI agents make possible and their actual uses and consequences.

From Concept to Practice: The Liminality of Mobile AI

As emerging technologies get onto the diffusion path, they progress from being completely new, or nascent, to a “liminal” stage of emergence (Ling, 2023). At this stage, a technology is just becoming familiar to the public, and there is a great deal of uncertainty about what kind of trajectory it will take in terms of uses and consequences. At the time of writing, mobile AI is in a liminal stage. In fact, many of the applications highlighted above are still nascent or even forthcoming. Just because agents are designed to optimize human tasks does not ensure they will have entirely positive social consequences. To illustrate, back-and-forth messaging to coordinate plans can entail more than mere drudgery. In some cases, the back-and-forth of microcoordination (without AI) can serve as a phatic thread of interaction that enriches the relationship. Scheduling and coordinating plans, particularly among close and romantic ties, can demonstrate commitment, while creating space for communicating about relationships and generating feelings of connection while apart (e.g., Ling, 2008). Mobile AI agents may offer efficiencies by offloading the back-and-forth, but this also might remove some of the “noise” that provides texture to human relationships (e.g., Licoppe, 2004). Furthermore, mobile AI may introduce and reinforce digital inequalities (Hargittai, 2003). During the liminal stage of diffusion, these types of unintended and negative consequences can begin to take root (Ling, 2023).

For these reasons, we cannot forecast how mobile AI agents will continue to be taken up and change the social ecology. Instead, we will continue to interrogate the possibilities for addressability of individuals, data, and both together, for a conceptual understanding of how ongoing developments in mobile AI relate to developments in mobile telephony and mobile broadband. Whether and how mobile AI represents a meaningful shift in MMC depend on how its liminality continues to play out. Rather than proposing it “is” a major development, this essay addresses how and why it could be, while recognizing the potential for unintended paths.