It's hard to talk about technology these days without AI entering into the discussion, and mobile app development is no exception. From Apple intelligence on an operating system level to chatbots becoming a part of so many applications, it's hard to find large technology companies these days that are not trying to put AI into their products and services.
Is this actually a useful addition to applications, and if so, do the positives outweigh the negatives?
There are many different types of AI being used. However, LLMs (Large language models) are the most common, these are the kind used for chatbots and support assistants you might have found (or have struggled getting an answer from) in banking apps, fitness apps such as Runa giving feedback on runs, or maybe you have chatGPT downloaded as an app straight to your phone for easy access to answers.
Although LLMS are the most common, there are many other interesting types that may offer more value. ML (machine learning) models can be used for creating personalised recommendations using behaviour analysis. Image manipulation can now be done with AI instead of traditional photoshop and many more.
These use cases can be advantageous if used in your app correctly. Chatbots that help answer the most common questions can be easily made, and the computation of statistics to give clear insights and feedback can be a really helpful addition to an application.Another use case for having AI integrated into mobile app development in giving you tailored experiences and advice, through access to personal data ai can help creating calendar updates from emails, setting up reminder lists for tasks it can see coming up and generally helping set up devices as a personal assistant. Increased automation of these everyday tasks through AI can allow users to have more free time.
There are many drawbacks to consider and balance when thinking about implementing artificial intelligence into an application. As with any feature, there are increased time and costs associated with the choice. Because AI systems are still relatively new, development can involve higher-than-normal expenses, particularly when specialist engineers are required to integrate machine-learning models, manage secure data pipelines, or build compliant AI-driven functionality.
There are also significant privacy and data-security considerations that must be addressed. Organisations must ensure that user data is protected, processed responsibly, and transmitted securely when interacting with large language models. This includes implementing robust encryption, enforcing access controls, following data-governance best practices, and ensuring compliance with privacy regulations such as GDPR. Clear policies around data handling, consent management, and secure storage are essential for maintaining user trust.
As AI usage has grown, we have also become increasingly aware of the environmental impact associated with running these resource-intensive models. Large-scale model training and inference can demand significant computational power, raising concerns around energy consumption, sustainability, and the broader ecological footprint of AI technologies.
Whilst adding AI can seem like the done thing at the moment, we think adding it in for the sake of saying your app is now "Powered by AI" is a risky game.
Implemented badly, these can work less well than traditional solutions and leave users frustrated chatting to a robot not getting the answers they were looking for.
Adding in AI when there is a genuine use can be beneficial, analysing data to give good feedback, editing images, sound or video in ways users would have to spend a long time learning on their own.
Some of the best uses might not be the ones given to the end consumer, using AI to help understand how users are using your application and allowing the developers to make better changes from feedback amassed and processed by the AI is one example of this.
Overall, AI, used with purpose, can elevate an app above its competition, but the pros may well not outweigh the negatives if thrown in for the sake of it.
