As AI agents grow more sophisticated, traditional mobile apps could face a wave of extinction, paving the way for an explosion of "OpenAI-style" applications. However, ensuring personal data security will emerge as a significant challenge in this transition, according to Hsi-peng Lu, a distinguished professor in the Department of Information Management at National Taiwan University of Science and Technology.
An AI agent is defined as a system or program capable of autonomously performing tasks on behalf of a user or another system. Many experts identify AI agents as the next major focus in the development of generative AI.
These agents serve as powerful personal AI assistants, automating tasks across multiple systems. To achieve greater automation, Lu stressed, technological upgrades are necessary for application programming interface (API) integration and cross-platform data processing.
Market research firm MarketsandMarkets projects the AI agent market to grow from US$5.1 billion in 2024 to US$47.1 billion by 2030. Advancements in natural language processing enable AI to better comprehend users' languages and dialects, allowing it to manage more complex queries and tasks. Additionally, the rising demand for customized AI solutions across various industries is fueling the growth of the AI agent market.
Lu employs the concept of ecosystem evolution to illustrate the impact of successive technological revolutions. In the 1990s, numerous standalone PC applications existed; however, the internet's arrival led to a mass extinction of these applications, replaced by a boom in online applications. The emergence of the iPhone then sparked a proliferation of mobile apps, drawing users' attention to mobile devices.
Now, the surge in AI agents driven by generative AI may trigger another mass extinction of mobile applications, potentially replacing them with interconnected "OpenAI-style" applications. Mature AI agents will be capable of performing cross-system tasks, such as information aggregation and utilizing user data for shopping assistance and restaurant reservations.
However, Lu cautioned that this increased convenience also heightens concerns about data security. Currently, user data—comprising personal information and usage behavior—may be dispersed across 100 different apps. With AI agents leading the charge, this data could be consolidated under the control of a single or a few systems.
Currently, mobile apps undergo specific security checks before they can be listed on app stores. But when it comes to enterprise-specific AI systems, who will conduct these checks? Will AI agent users' identities be vulnerable to impersonation, and how should authentication be managed? These concerns are prompting many companies to develop their own language models.
Rock Tsai, chief information officer at Taiwan Mobile, noted that IT development is currently in its fifth major cycle. The previous cycle, centered around mobile devices, lasted from approximately 2000 to 2020.
As AI technology continues to advance and agent-based applications become more prevalent, the ways businesses interact with customers and among themselves may undergo significant changes. Tsai believes that AI will remain a driving force for development for at least the next decade, urging businesses to consider how to integrate AI applications into their existing digital transformation strategies.
Article translated by Willis Ke
