ChatGPT for future trend forecasting
A fast-evolving cultural landscape has always compelled future forecasters to seek novel ways of harvesting future trend signals. The advent of Chat GPT has led many forecasters in the world to ask themselves: how can I use AI to boost my work?
This is what I discussed in my monthly column for Spur Magazine's August print issue.
In the past year there have been many guides on how to use Chat GPT, with varying degrees of quality. Some created by genuine experts, some spun out to monetise on the virality of all things AI.
Prompt engineering, sometimes dubbed ‘prompt whispering’ is the skill to now master. Open-source community platforms such as Learn Prompting offer comprehensive trainings. Daniel Meissler’s Unsupervised Learning also provides information from the philosophical implications of AI, prompting, coding to issues of security.
But let’s be clear. Chat GPT can’t predict future trends. However, should you want to solely focus on a short term trend analysis (6 months to a year), gathering impressive summaries of large data sets, Chat GPT has forever facilitated your work.
Chat GPT is relevant especially for summarising secondary future trend research as it pulls answers from everything ever recorded on the internet. It can help in a variety of data insights. From analysing trends based on large amounts of fashion-related language data such as fashion blogs, social media posts, and fashion magazines, to identifying key influencers and the trends that they are promoting. Chat GPT can also be used to predict consumer preferences by analysing language data from social media and online shopping sites.
For generating new design ideas, Chat GPT can lubricate the early design process by analysing language data from fashion-related sources and generating ideas for new styles and designs. When it comes to AI generated visual designs, large language models offer up impressive results, as seen with G-Star’s recent AI generated denim couture piece. This shows great promise for visualising and prototyping anything from campaigns to designs and possible future scenarios.
When it comes to pulling data, writing facilitation, producing summaries, the integration of Chat GPT by Google, Notion, Microsoft 365, and many other platforms indicates that this technology is infiltrating our daily productivity suite, with the likelihood of humanoid robots becoming a work companion. The possibilities of AI and Artificial general intelligence (AGI) are dizzying.
Chat GPT has certainly saved me time in summarising ideas. But as a futurist exploring cultural theory and how this impacts the future of fashion, it remains a tool, not an end.
Albeit a fascinating new and shiny tool, it offers no ingenuous understanding of the future, especially future macro trends. This is because there are many external factors and inherent uncertainties that can influence the complex trajectory of trends over longer periods of time. The other missing factor is Chat GPT’s inability to incorporate a philosophical, ethical and creative view of possible futures.
Outsourcing the creative and philosophical part of future foresight work to a Chat GPT is a worthy experiment, one that Nemesis trialled by writing with Chat-GPT3 in their DOOM! Report. But it remains essential for humans to cultivate critical and creative thinking, to conjure the most cutting edge interpretations of the future. Without these skills of fine-tuning, forecasters run risks we don’t yet fully understand.
An obvious risk is that whilst Chat GPT can help with our knowledge work and data insights, most of the answers are things that are already increasing in demand, and therefore lack in originality.
What is also lacking in the way many people are thinking about A.I. is the idea that knowledge work is this type of systematic task you must fit into an ultimate productivity model. And now to make your life quicker, Chat GPT can spit out the output for you. But what about the mystery in the process of future forecasting?
For instance, in the work of filtering trend signals, a first draft, even if bad, of future trend concepts is not a waste of time. It is often an intellectual exercise, an 'in between' space for cultural theorising that we need to spend time in. This draft space leads to creative foresight. The idea that we can just outsource this 'unquantifiable' moment means we are thinking of ourselves as computers as opposed to the more magical creatures we are.
Another risk to consider is the inherent biases in generative AI applications, as these systems reflect human biases that shape results around conventional societal stereotypes in gender, age, race, and nationality.
Certainly, by using Chat GPT for future trend forecasting some things are to be gained. Chat GPT summarizes essential data. And the spirit of exploration should always come first for a future trend forecaster.
(This next part wasn't in my original piece for Spur but I am adding for further reflection.)
The mass proliferation of AI and large language models is a paradigm shift, and that's an understatement. Chat GPT and its counterparts are just the visible tip of the iceberg.
Here are my main takeaways so far when it comes to using ChatGPT for future trend forecasting:
Instead of generating 10 ideas a minutes, AI can generate 100 a second. Is this speed of output what we need in future trend foresight to generate great ideas?
We should aim to combine human originality with the efficiency of smart generative machines.
For a human being to use AI effectively, given it still produces mistakes, it is necessary to have a solid base of general and sector-specific knowledge.
Investing in our philosophical and creative learning has never been so essential
How will we guard our cognitive abilities from further distraction and go 'beyond the matrix theory of the mind' to quote Ezra Klein?
Due to the uncertainty inherent in trends, we have to cultivate contemplation and creativity, ChatGPT or not
Read the full column in Spur Magazine August’s printed issue out now.
By Geraldine Wharry
