Can human-centricity and AI co-exist? We asked the AI

With ChatGPT, artificial intelligence (AI) can now produce grammatically correct sentences, with  coherent storylines. Everyone agrees this will have a huge impact on industries and people, but nobody knows exactly how.

Meanwhile, the human side of design has come a long way. Can and will designers now have to hand over the baton to AI? Or how will human-centric design be affected with the emergence of tools of ChatGPT’s caliber?

To get an indication, who better to ask than the ChatGPT itself? The following is our short interview. The answers were to the point, intelligent, well phrased — and therefore scary. We have abbreviated some of the answers since, as it turns out, he/she/it is quite chatty.

As an aside, it is interesting to consider how ChatGPT works. The GPT part of the name — meaning Generative Pre-trained Transformer — is a type of language model developed by OpenAI. It is geared towards discussion-like interaction, but there are other language and non-language models. 

So the AI we’re talking to here is just one iteration. This will undoubtedly inspire endless variants we cannot even dream of today, of course also “inspiring” interesting ethical discussions at the intersection of designers, users and AI. Exciting times ahead! 

But now, let’s see what we’re working with here. Over to you, ChatGPT.

AI in human-centric disciplines 

Question: “How do you see human-centric design and AI co-evolve?”

ChatGPT's answer: 

Human-centric design and AI are likely to co-evolve in a number of ways. One way is that AI will be used to augment the design process, allowing designers to generate and test multiple design options quickly and efficiently. 

Additionally, AI will be used to personalize and optimize products and services for individual users, based on their preferences and behaviors. In the future, AI may even be able to take on some of the tasks currently performed by human designers, such as creating 3D models or generating design concepts. 

However, it's important to ensure that the AI system is developed with a human-centered approach, taking into account ethical considerations such as privacy and bias.

My reflection:

I agree. It is a better answer than I could have given. And there are caveats that we must understand:

1. All pre-trained models are blind to bias and errors in training data. Existing models can be as good as our best understanding of the area to be modeled.
2. Bias, privacy, and errors in model data are not something the companies developing or using AI want to emphasize. As legislation in all countries is lacking in the area of AI, we should demand ethical and legal clarifications on how AI may (and may not) be used, and how end users need to be informed about the use of AI. (Just… please don’t make it another “do you accept all cookies”-popup solution.)
3. As with all automation, there will be an impact on the job market. The chatty AI will take jobs this year, next year even more. This will need difficult political answers also.

AI for UX designers – or against?

Question: “How can AI help us to get rid of user interfaces?”

ChatGPT's answer: 

AI can help to reduce the need for user interfaces by allowing for more natural and intuitive forms of interaction between humans and technology. Some examples of this include:

• Voice assistants
• Gesture recognition
• Virtual and augmented reality
• Proactive AI: AI can be used to anticipate the user's needs and perform actions automatically, without the need for explicit input from the user.

My reflection: 

This all is true. The last one is also scary. As mentioned earlier, current models can, at best, only be as good as our understanding. There is bias in the models, and incorporating that bias into actions people take, may easily increase the bias.

To put it bluntly, we are not yet ready for proactive automation. The ethics must come first. On some tasks – like gesture recognition – AI may help a lot. Even gesture recognition is not devoid of bias (“was that person showing me “OK” or a “your ego is of this size”-gesture?”), but the models for computer system gestures are much better understood than the use of gestures between humans.

Question: “Please tell me what relationship humans and AI will have in the year 2038.”

ChatGPT's answer: 

It is difficult to predict exactly what the relationship between humans and AI will be in year 2038, as it will depend on a variety of factors including technological advancements and societal attitudes towards AI. However, it's likely that AI will play an increasingly large role in many areas of life, including transportation, healthcare, education, and the workplace.

In transportation, it's likely that self-driving cars will be widely adopted, which could greatly reduce the number of accidents caused by human error. In healthcare, AI may be used to assist doctors in diagnosing and treating diseases, and in education, AI-powered tutors may be able to provide personalized instruction to students.

In the workplace, AI may be used to automate repetitive tasks, and to help managers make data-driven decisions. This could lead to increased productivity, but also to job displacement.

It's also likely that the ethical and legal issues around AI will become more pressing, as the technology becomes more advanced and more integrated into society. There will be a need to ensure that AI systems are developed and used in a way that is fair, transparent, and respects people's privacy and rights.

Overall, the relationship between humans and AI in 2038 will be complex and multifaceted, with both benefits and challenges. It will be important for society to continue to engage in thoughtful and ongoing dialogue about the role of AI in our lives and to work towards creating a future where humans and AI can coexist and thrive together.

My reflection: 

… yeah, it is good.