ChatGPT is making all the headlines these days, and beyond impacting organisations and consumers at large, it is starting to have a significant impact on and raising questions across the developer community.
Why is that so important? Because developers are the ones creating the apps, software and digital products that are the fabric of our digital world, and what most Australians organisations today rely upon to operate.
Savvy end users today expect a certain level of customisation throughout their experience in an app, website or platform, and Artificial Intelligence (AI) tools such as ChatGPT – which offers an AI service in crafting written responses with an input of information as the base – alleviate some of the stresses developers face when building these digital products.
But there are important limitations to using AI for a convenient but effective developer experience (DX) that heavily digitally reliant organsations should pay attention to.
Is every AI a good AI?
As end users continue to expect more customisation, efficiency, convenience and personalisation, businesses don’t have a choice but to embrace AI to remain competitive.
This is what has driven the incredible popularity of AI platforms like ChatGPT which make developing these functionalities much easier.
It is very tempting for organisations to throw dollars at this kind of technology for all the aforementioned advantages it offers, but it’s equally important to understand how sophisticated AI’s machine learning can impact the end user’s journey.
For example, if a customer can’t tell if they are speaking to a human or a bot while opening a chatbot function, they might move to another company or product that includes more interactions with easily identifiable humans.
The bottom line is, AI cannot replicate a real person’s tone or even real-life experiences, it is not a replacement for live experience or built personal knowledge.
The human side of software: keeping connections and empathy alive
Our interactions have evolved to digital first in many cases and because of this, organisations and their developer teams should ask themselves:
- Is the digital experience being created the best it can be?
- Is the digital-first strategy human-centric?
- Would the user experience really be better if for example a chatbot were enabled, or would core human interaction elements be lost?
Using AI shouldn’t be to the detriment of some core human interaction qualities, or lessen opportunities for deeper engagement.
With a better, deeper understanding of the end user and their experiences, their needs, their reason for engaging, and by communicating with users, it is possible to build software that has empathy qualities and respects the nature of human connections.
Focus should particularly be put on:
- Using user feedback, user analytics, competitive analysis, market research and more to understand what users need and value as humans
- The app or digital product should allow for clear and effective communication
- Privacy and security should be front and centre, with strong data protection measures in place and transparency around how user data is collected, used and shared
- The social impact of replacing in-person engagement with AI-driven apps should be considered: will the solution promote social connections and engagements, or will it contribute to isolation and disconnection?
- Accessibility – will the digital service or product be accessible to a diverse range of users, or will it reinforce existing social inequalities?
Tools like ChatGPT, while extremely useful, are ultimately machines and cannot truly understand or feel emotions in the way that humans do. They can simulate empathy to some degree, but it is important to remember their limitations and use them to support human empathy and understanding, rather than a replacement for it.
Let’s not forget the AI bias issue
AI bias has been discussed for about a decade now, but tools like ChatGPT mean it is even more important that we train our models without bias – and this means being self-aware.
Organisations wanting to capitalise on the power of AI need to make sure they use unbiased data and algorithms as it allows them to avoid perpetuating or reinforcing existing biases.
This can include carefully selecting training data and ensuring that algorithms are transparent and explainable.
The rise of this ChatGPT era means developers now have to walk a very fine AI line and understand what can be helpful and what can hurt the development of tomorrow’s apps and software. Shiny and new doesn’t always mean the best option for developers. And while competitiveness is something to consider for modern developer experience platforms (DXPs), it’s just as important to understand human audiences and how they might perceive AI that will likely be their first introduction to a service or platform.
AI does have a very real and strategic place in software. The onus is on the developer to harness the technology in the best way possible while still keeping the human end-user experience as a top consideration.
Sara Faatz, Director, Technology Community Relations, Progress.
