Skip to content

Generative AI: Understanding Fears & Embracing The Innovations 

Now is the time push away fears and embrace artificial intelligence (AI) technologies.

TW Special Report

For some, the mere idea of Artificial Intelligence (AI) conjures up fears of robots and computers taking over the world in some future dystopian nightmare. This futuristic idea seems less and less far-fetched as generative AI (GenAI) technologies —a subset of AI that creates text, images, videos or other forms of content based on data it has learned — are starting to take off. GenAI often produces output as a response to specific prompts, or requests, given by a user, with Chat-GPT one example of such a technology.

But in the shorter term, perhaps the most common fear expressed is a more basic, “Will AI take my job?” Half of regular AI users think their job may disappear within the next decade according to a report, “AI at Work: Friend and Foe,” released by the Boston-based Boston Consulting Group. The company’s BCG X Tech Build and Design Division surveyed 13,000 employees in 15 countries and regions about AI use and adoption.

“Our survey exposes the double-edged nature of GenAI,” said the report’s co-author Sylvain Duranton, a managing director and senior partner at BCG. “Familiarity correlates with both comfort and fear. GenAI is a revolutionary technology, so these opposing reactions should not be surprising. By recognizing the complex ways in which humans understand and interact with GenAI, leaders can reshape their organizations to maximize the strengths and value of both their human and machine workers.”

BCG’s survey also revealed that engagement with AI in the work-place is increasing. Almost two-thirds of leaders report they are beginning to use GenAI tools in their organizations and more than twice as many frontline employees reported using AI tools on a regular basis this year compared to 2023.

AAPN Embraces AI, Hosts Workshop

Joe Parrish, founder of Winston-Salem, N.C.-based award-winning ad agency The Variable Agency, doesn’t think AI will take jobs. “However, a big qualifier to that is a person with AI experience may take your job,” Parrish conjectured during his presentation “Leveraging AI In Your Business— Apparel or Otherwise,” given at an AI Executive Workshop hosted recently by the Atlanta-based Americas Apparel Producers’ Network (AAPN) for its members. AAPN is a huge proponent of using AI and wants its members to embrace the technology sooner rather than later (See sidebar).

Parrish views AI as a transformative technology, with the potential to be more impactful than steam power was during the industrial revolution. “If programmers were to stop developing the current AI technologies tomorrow, I predict it could still take a good 10 years to fully realize the capabilities of the existing technology,” he said. “We have tons of capability at our fingertips, but we have a utility problem. Generally, people don’t know how to use AI. But I truly believe there isn’t a problem that AI as it exists today couldn’t help us solve.”

Fast-Changing Technology

AI technology is evolving rapidly, and to illustrate the rapid pace of change, Parrish introduced Chat-GPT’s brand new voice feature to the audience during his presentation.

Because of this rapidly changing technology landscape, Parrish recommends jumping in and using AI now. Start small, and grow with the technology. “The best time to dive into AI was yesterday.” But the good news according to Parrish is that the second-best time to jump in is today. “There will never be a smaller gap between an AI beginner and an AI expert than there is right now,” Parrish added.

Transformative Technology

AI may be used in a variety of ways in industry, academia and even one’s personal life. Some of the applications include image generation and recognition, language translation and processing, machine learning, computer vision, and decision making — analyzing data to find patterns and offer insights to aid informed decisions — among many, many applications for AI.

Ethan R. Mollick, associate professor at the University of Pennsylvania’s Wharton School, collaborated with a team of social scientists at BCG on an experiment to assess the future of professional work in the age of AI. Mollick shared that when consultants were given 18 different tasks designed as realistic examples of the type of work performed at elite consulting companies, “… consultants using ChatGPT-4 outperformed those who did not, by a lot,” Mollick said. “On every dimension. Every way we measured performance.”

Specifically, consultants using AI completed 12 percent more tasks on average, were 25 percent faster and produced 40 percent higher quality results than consultants not using AI.

As a comparison, steam power increased productivity by approximately 18 to 22 percent. And today, companies spend huge amounts of money on massive software installations to see perhaps a 3 to 4 percent increase in productivity. In studies such as the one from The Wharton School, AI is exceeding these productivity gains. By a lot.

Parrish recommends assuming AI can help whatever the situation. “Try it, and let it prove you wrong,” he said. He notes it can take at least 10 hours to understand what the technology can do and how it can help. “But too many people try using AI a few times, don’t like what they see and give up,” Parrish said. “Use AI as a thought partner. Let AI take on tasks that take you lots of time, sort of like a grad school intern.”

He also suggests taking work tasks that are time consuming and creating a ChatGPT to solve the problem or reduce the time needed for a task.

What Type Of AI

Two of the biggest AI breakthroughs, according to Parrish, are ChatGPT and Midjourney. “There are other facets to AI, but these are more expensive and more specific,” Parrish said. “Chat GPT and Midjourney are tools everyone can see value in.”

During the AAPN workshop, Parrish talked in-depth about both technologies and how and when to use them, among other AI technology options (See Figure 1).

Different AI technologies have certain strengths, and the ultimate end result may be best generated using multiple GenAI technologies. For example, “ChatGPT 4o is fairly terrible at generating images, but is great for creating prompts to give to Mid-journey,” Parrish said. According to Parrish, Midjourney is one of the best AI image generation tools to choose. But a user can utilize ChatGPT to create a prompt for an image, then take that prompt over to Midjourney to obtain a different, better result.

According to Parrish, ChatGPT 01 is very good at math as well as thought and reasoning, and step-by-step problem solving. Claude AI is a great all round Gen AI tool; Complexity is adept for research; and Google’s NotebookLM is great at condensing and summarizing lots of information into study guides, FAQs, timelines or even a podcast you can listen to, among other ideas.

GenAI also is popping up in software that is already commonly used. Gemini is Google’s AI. It may be used to write text or to create images within a Google slides presentation, for example. Microsoft Edge also now includes a button for Copilot — the name for its GenAI technology — in the browser. Save time reading a long article and use Copilot to summarize a story, for example.

Some GenAI technologies are free, or have a free version, while others are paid only or offer additional services with a paid subscription. Most are available with low monthly fees or team/group rates for a company wishing to offer access to all its employees.

Using AI: How To Engineer A Prompt

Engineering a prompt for GenAI is a skill that needs to be developed. AI is intelligent, but the wrong prompt may not generate the answer that is sought. An individual’s value comes from the ability to write the correct prompt. “No two people will get the same result from AI, because the inputs are not the same,” Parrish said. In addition, the more a person can incorporate expertise from their vertical, the better a prompt will be.

To start, break down any job into a series of smaller tasks. AI cannot write an entire movie, but it can generate ideas for a scene in a movie, for example.

Parrish introduced an analytical method of writing a prompt he defined using the RTEF acronym, which stands for Role, Task, Example and Format. The mnemonic “Robots Teach Everything Funny” is a helpful way to remember these four prompt requirements.

  • Role — who. Give AI a role —teacher, engineer or writer, for example.
  • Task — what is the problem, what is needed?
  • Example — perhaps suggest some examples to consider.
  • Format — state the desired out-put format. For example, “Give me a title for the story and a one paragraph summary.”

Prompts also may simply be conversational in nature. Conversational prompting lets you build expertise by uncovering insights gradually, adapting questions based on responses. “It’s a longer process that deepens your understanding over time, allowing you to refine your knowledge through each interaction,” Parrish said.

Parrish breaks down Midjourney prompts for images into three levels of Atomic Prompting. The first, or organism level, level is a basic image prompting consisting of an art direction and the subject of your prompt. The second, or molecular level, is a more detailed approach to prompting that includes descriptors like inspiration and shot type for reference. The third, or atomic, prompt is a highly specific prompt to control every aspect of your image from environment to styling. He demonstrated the quality of the images generated as he refined and perfected his Midjourney prompts during the AAPN AI workshop.

If Midjourney isn’t producing what you are looking for, reprocess the prompt. Parrish showed that Midjourney will produce images that are very different given the same prompt (See Figure 2).

AI Misconceptions

AI is generally viewed as an analytical tool that is not creative. “But it’s not true,” Parrish said. “If AI is not creative, that’s because the user is not giving it the right prompt.” To illustrate this, he noted that the best test is an alternative uses test.

When asked for ideas of things a pencil can do other than its basic function of writing, on the low end, humans can come up with three to five mostly obvious uses. On average humans generate between 5 and 10 ideas, while a few exceptional people specifically trained in divergent thinking may manage to exceed 15 ideas. When asked, AI suggested 100 alternate uses, far surpassing humans in this creative task. “Given the right prompt, AI can be very creative,” Parrish said.

Perhaps this may not appear to be creative in a human sense, however, answering a prompt decisively is AI’s relentless mission.

Specific Apparel Industry Ways To Use AI

During the workshop, Parrish spent some time explaining popular AI tools, showing participants ways to exploit the technologies in various aspects of the apparel industry.

AI can be used to create trend presentations, mood boards and even apparel designs. Parrish created a theoretical apparel collection using a variety of AI tools all the way to photographic-like images of a model wearing the designs (See Figure 3).

Ad agency executive Parrish also demonstrated using AI to develop a new product including brand name, logo, related photographic images, a song generated using Suno or udio AI, and finally a video to showcase just some of the capabilities of GenAI. “AI can create commercials from start to finish and it’s only getting better,” Parrish said.

Analyzing Data

ChatGPT can handle the mammoth task of analyzing data sets such as shirt sales for the year, for example. Once the data is uploaded, ask ChatGPT questions like: “What are the different ways I can analyze this data,” “Let’s do a comprehensive analysis of stock turnover and back orders,” or “I have to prepare a report using this data. Outline the report and create the charts I will need to look great in front of my boss.” AI can examine the numbers and see patterns, trends and outliers quickly and effectively.

It may feel daunting to share large, private data sets with AI technologies, but the data is safe. AI uses the data to learn and analyze, but does not share this information with anyone else.

Getting Started

Bringing a company into the AI era may at first seem daunting. But apply the proverbial “eating the elephant” idea and start with small “bites” applying AI to small tasks. As a company or an individual, breaking down a collective fear of AI may ease the transition and spark a broader corporate adoption.

Parrish suggests companies can think about the following five things to encourage the adoption of AI:

  • Cultivate curiosity — In the age of AI, foster a culture of continuous questioning and exploration. Employees should be encouraged to ask “what if” and “why not,” pushing boundaries and seeking new applications for AI rather than settling for the status quo.
  • Reward innovation — Recognize and incentivize employees who propose novel ways to leverage AI, even if not all ideas are implemented. This encourages a culture of forward-thinking and helps the organization stay ahead of AI trends.
  • Focus on augmenting, not replacing, humans — Frame AI adoption as a way to enhance human capabilities rather than a threat to jobs. This approach not only eases adoption but also leads to more effective human-AI collaboration, leveraging the strengths of both.
  • Prioritize adaptability —AI technologies evolve rapidly. Companies need to build flexibility into their workflows and be ready to pivot quickly. This means valuing employees who can learn and adapt swiftly, and creating systems that can be easily modified as new AI capabilities emerge.
  • Value skills that complement, not compete, with AI — Rather than focusing solely on technical skills that AI might replicate, organizations should prize uniquely human abilities like emotional intelligence, creative problem-solving, and ethical decision-making. These skills work in tandem with AI to produce superior outcomes.

Parrish suggest individuals consider the following four things related to AI:

  • Value the impact, not work hours — Working hard is important — it offers the experience to learn and become an expert at something. But once a task is learned, the path to scaling impact without scaling work hours is significantly clearer thanks to AI. When deciding what to work on, it’s increasingly important to find the areas that create the most value.
  • Assume the skills to solve the problem are present — Consider learning something new from ChatGPT. It may be surprising how AI cuts the time required to learn a new skill down to a fraction, or even a fraction of a fraction, of previous efforts. With the explosion of content online and the accessibility that AI offers, learning a new skill and solving a problem has never been easier.”
  • Be resourceful — When resourcefulness is present, AI serves as the ultimate resource. Thanks to AI, the primary limitation to achieving dreams is likely internal. People are launching software companies without prior coding experience. What accomplishments could AI support?”
  • Specialize and dig in deep —Expertise is rising in value as AI commoditizes average. By design, AI generates the average of human intelligence on any task. Being a generalist may become easier, leading to an average skill level in everything with minimal effort. This trend will likely apply universally.

Leveraging AI For The Future

“We are entering a new era for GenAI which is less about optimism and curiosity and more about confidence and value realization,” said Vinciane Beauchene, a managing director and partner at BCG and a coauthor of the “AI at Work: Friend and Foe” report. “Adoption has increased, and individuals are starting to see the benefits. Companies are also starting to realize that get-ting the value out of their investment will require them to think beyond productivity and take a more holistic and proactive approach to redirect the time saved to the most valuable and joyful activities, to reskill their employees to do so, and to reshape their organizations and operating models as a consequence.”

Some companies in the textile industry are already embracing the use of AI in their day-to-day operations such as World Emblem, an embroidered patch manufacturer based in Fort Lauderdale, Fla. (See “World Emblem: Embracing AI,” TW, this issue). And AAPN hopes many more textile companies will take the leap and at least dabble to get more comfortable with the idea of using AI in their businesses.

And maybe, just maybe, GenAI contributed to this article.

The Americas Apparel Producers’ Network (AAPN) is an Atlanta-based association serving the apparel industry supply chain “from the dirt to the shirt”. Its member companies are located in North America, the Caribbean, Central and South America, as well as Europe and Asia. AAPN believes sourcing in this hemisphere makes sense — it’s easier, faster, safer and more stable, among other benefits.

AAPN is looking at ways to integrate and use AI in the apparel industry. It wants its members to embrace, implement and use AI to its benefit. To AAPN, it’s not about taking jobs away, but about making them better.

“AI represents a pivotal moment for our membership, many of whom are leaders in the textile and apparel industry,” said Lynsey Jones, executive director, AAPN. “At the AAPN, we view AI as a powerful tool that can provide a competitive edge, helping our members further solidify their leader-ship positions. Whether it’s automating routine tasks to free up time for more meaningful work or leveraging AI to enhance customer experiences and support data-driven decision-making, the potential is vast. Importantly, we don’t see AI as a replacement for people, but as a catalyst for productivity, creating opportunities for members to focus on higher-impact initiatives.”

Jones scheduled a call with Joe Parrish and his team after a conversation with an AAPN member who had engaged him to walk their company through the AI revolution. “From my very first conversation with the team, I was captivated,” Jones said. “Their enthusiasm for AI and its practical applications was contagious. Joe made something as complex and daunting as AI feel accessible and approachable.”

AAPN initially brought Parrish in as a speaker at its 2024 pro:Americas Annual Conference. “The entire room was just as impressed and inspired as I had been during that initial call,” Jones noted. The feedback and response prompted AAPN to put together the AI workshop to continue the learning.

Editor’s Note: This feature is based on information presented by Joe Parrish at the Americas Apparel Producers’ Network (AAPN) AI Executive Workshop— “Leveraging AI In Your Business —Apparel or Otherwise.” Parrish is the founder of The Variable Agency in Winston-Salem, N.C.

2024 Quarterly Issue IV

 

Leave a Reply

Your email address will not be published. Required fields are marked *