Transforming Interactions with ChatGPT
When engaging with a chatbot, users often encounter lengthy text responses, sometimes accompanied by one or two images sourced from the internet. However, OpenAI is revolutionising the way ChatGPT interacts with users. On Wednesday, the company unveiled a comprehensive update to the “intelligent user interface” of ChatGPT, designed to generate customised visual elements that enhance the responses to user queries. This new experience is powered by the recently launched GPT-6 model, which became available to paid users yesterday and is now accessible to free users as well.
An Insightful User Experience
I had the opportunity to briefly test the new user interface prior to its official release, and the visual elements introduced represented a significant improvement. For instance, when I requested a diagram of a slug, ChatGPT produced an image that focused on its anatomy, complete with buttons that could be clicked to explore each part, from the upper tentacles to the muscular foot. Additionally, I was able to create an affordability calculator for apartments in San Francisco, which featured sliders to adjust income and other expenses, clearly indicating what percentage of my net salary various housing options would cost.
One of the standout features during my brief trial with ChatGPT’s intelligent user interface was the airplane seat explorer. As I am not fond of sitting in exit rows, I asked ChatGPT to generate a tailored tool to compare cabin layouts across four different Alaska and Delta aircraft. The outcome allowed me to click on any seat within the map for further details, with exit rows highlighted in green.
Visuals When They Matter
While these visual enhancements prove beneficial in certain scenarios, not every response from ChatGPT suddenly features a chart or table. Aarush Selvan, a product manager at OpenAI, explained, “Our design team has invested considerable time determining when to incorporate a diagram, chart, or buttons to add value, and when it may feel cluttered.” If a user prefers fewer visual elements, they can simply instruct ChatGPT to produce less of that nature.
Following in Google’s Footsteps
These visual outputs bear resemblance to what Google announced earlier this year for its search engine, expanding the personalised results available to users. Google refers to this as the “generative user interface.” For instance, if one wishes to gain a deeper understanding of deep space, the search engine might display an adjustable graph that breaks down how black holes function, utilising various sliders for interaction.
Embracing the Multimodal Era
Interactions with chatbots are increasingly becoming multimodal, integrating text, images, and voice into a cohesive user experience. OpenAI’s launch of an updated version of ChatGPT, which is visually appealing and more interactive for over a billion users, signifies that the multimodal era is upon us. As the shift towards generative user interfaces continues, AI is progressively dictating the appearance and functionality of software tailored for individual users.
The Future of Personalised AI
As AI tools become adept at swiftly generating bespoke software, the future of chatbots promises to be even more personalised and aligned with user preferences. Personally, I look forward to seeing more slug diagrams rather than endless walls of text. The evolution of chatbot interactions is undeniably exciting, marking a departure from traditional formats and ushering in a new era of engaging communication.