Recapping NCSU’s AI in Textiles and Fashion Conference

I visited North Carolina State University last week to attend their first ever “AI in the Textiles & Fashion Industries” conference.

The event was a good mix of showcasing some research being done at the Wilson College of Textiles, identifying areas where AI will be impactful and panel discussions with people from industry.

Much of the discussion reiterated familiar themes, including:

• The need for guardrails as AI continues to evolve.

• Data quality used to train AI models needs to be impeccable.

• AI is about enhancing operational efficiency, but suitable use cases are needed to get necessary funds to develop it.

• People are resistant to change.

What is also true is that the human touch is as important as ever as AI develops. As we have argued before, advances in technology require a compensatory balance of the human touch.

This conference helped to punctuate how important the human touch is. For instance, graduate students doing fundamental research, collecting data points in the lab to train an AI model. Often, collecting a single data point can take a few hours. Having been a graduate student myself working in a lab, I totally understand the effort being undertaken.

This goes back to the data quality issue mentioned above.

Who is going to ensure that the data is impeccable?

The humans, of course.

Turning to the more specific applications of AI, there was much discussion about uses such as:

• Detecting textile defects.

• Dyeing/finishing e.g. predicting color fastness on a wet vs. dry textile.

• “On loom” fabric inspections in real-time.

• Development of a textile color data bank.

We have discussed previously the need for process innovation for brands. One example is a “materials-first” approach where brands design into already-available materials.

This might be an area where AI would be highly valuable. Being able to identify defects in the material as early as possible or detecting defects in the dyeing/finishing process is critical. Whether a brand buys into materials upfront before the season OR if they wish to buy materials “on demand” for products that are intended to be fast-tracked to market.

As such, digital transformation for materials would almost be incomplete without the use of AI.

At the end of the meeting, we were shuttled up to the Wilson College of Textiles for a tour.

We were shown the labs for knitting, spinning, dyeing/finishing and more. The space is absolutely impressive.

The students, staff and faculty that led the tour had an unmistakable and genuine enthusiasm for their work. I’m not sure how much I could be excited about weaving, stitching, the intricacies of a Jacquard machine and spinning fibers into yarn. But, everyone who I met has passion for their work in spades and it was wonderful to see.

If there is one key takeaway from my time at NC State, it would be this:

I don’t care how much AI advances, it’s a given that it will.

The human touch, and how much that advances in parallel, is just as important.

….if not more so.