Generative AI and ChatGPT: How the Tech Could Overhaul the Consumer Goods Industry and Red Flags to Look For
Although still an emerging area of interest in the consumer goods industry, generative artificial intelligence (AI) has made a prominent stance as tech with enormous potential and various applications, garnering a lot of attention, but also collecting several questions, concerns, and doubts.
To gain a better understanding of how generative AI might transform various fields across the industry, we’ve tapped into expert voices and research from Coresight, Microsoft, and Gartner.
While the consensus clearly points toward significant advancements in the area in the years to come, sources also agree that the way forward must be met with caution and small, strategic steps.
“Generative AI can result in biased and insensitive content based on the data it got trained on, as it needs a large quantity of data to train its AI models, which can damage the brand’s reputation,” warns Subroto Mukherjee, Microsoft’s strategy lead for cross industry solutions, metaverse, Web 3.0, NFT, AI, OpenAI, ChatGPT, gaming, and commerce.
Editor’s Note: Subroto Mukherjee’s contributions are his own viewpoints based on his personal experience, research, and learnings and do not necessarily reflect the opinions of his company.
What Is Generative AI?
Early iterations of artificial intelligence have been around since the 1950s. What then are the differences between general AI and generative AI?
Primarily, says John Harmon, senior analyst at Coresight Research, generative AI tech like ChatGPT uses deep learning neural nets that mimic brain cells to generate the best text from the inputs.
Meanwhile, other AI-powered tech like machine learning takes in data sets to identify relationships among data and make predictions.
“Many of these are quantitative forecasts/predictions, and the AI element ensures that the models evolve over time to stay accurate,” says Harmon.
- Generative AI History and Investment
Before launching into use cases, it’s important to identify that generative AI, while not abruptly new, has recently seen a surge in adoption as companies feel confident enough in the tech’s ability to release it for widespread use. But its foundation is wholly different from general AI in functionality, and models vary widely from each other.
“Technologies like neural nets have been around for decades,” says Harmon. “Many companies developed technology like ChatGPT a while ago, but didn't release it because it wasn't ready for a broad audience.”
The tech continues to evolve at a rapid pace, he says — largely due to healthy VC investment and an increasing amount of computing power which can enable new AI capabilities.
The global generative AI market size accounted for $7.9 billion in 2021 and is projected to occupy a market size of $110.8 billion by 2030, according to Acumen Research and Consulting, a global provider of market intelligence. And total private funding in China alone was estimated at $17.21 billion in 2021.
Generative AI Applications
Despite being so new to the industry, CGT predicts long-term benefits. In fact, we foresee generative AI playing a forward role in the consumer goods jobs of the future, particularly in the areas of upskilling. According to a recent Gartner blog post, generative AI will likely be used “to replace, recalibrate, and redefine some of the activities and tasks included in various jobs.”
Among its capabilities, said Gartner, are summarizing text, classifying content, answering questions, and translating and converting language (including programming languages).
Using these abilities as a guide, it’s safe to presume the technology can impact areas of business like marketing and e-commerce in order to provide more targeted marketing, elevated customer support, and increased personalization efforts to boost loyalty.
The top potential use cases for the consumer goods space, says Mukherjee, are “generating novel, diverse, and personalized copy content, audio, video, and images at speed and scale; and saving time and cost for audiences in different languages spread across other geographical boundaries.”
Not only can this lead to improved product packaging, he says, but can result in faster A/B testing and improved marketing performance, and can even influence some of the growing e-commerce customer support technologies like chatbots, virtual assistants, and avatars.
“They can provide real-time multilingual customer support and product demos, and any queries can be responded to based on customer sentiment in different regions resulting in improved customer satisfaction and reduced support costs.”
Across product innovation, generative AI can also provide more accurate recommendations, says Mukherjee. “It can scout for trends, social media data, tweets, sentiments, and features based on what customers are currently looking for.”
Harmon believes artificial intelligence could significantly transform three primary areas of business in the consumer goods space:
- Merchandising: Generative AI can quickly generate product descriptions and stories
- Content Creation: Meta, for example, has demonstrated the ability of AI to generate images and landscapes from voice commands
- Consumer Engagement: AI/ML can be used to find the language that generates the greatest response from consumers
Within these areas, CGT sees enormous potential, such as using image conversion tech within the beauty industry to streamline clunky AR applications, creating personalized product regiments by leveraging consumers’ past histories, and further tapping into consumer engagement efforts with generative AI-created quizzes and product education.
It’s easy to see this translating to several other categories, and expanding into even more use cases, but the key, according to experts, is to take it slow.
Keeping a Healthy Dose of Skepticism
Why the caution sign? With a technology that is reliant on data inputs, there’s always risk. And in the case of generative AI, these risks include copyright infringement, insensitive content creation, misuse of content, and brand reputation repercussions.
Take the virtual chat example as noted above: According to Mukherjee, chat agent conversations based on generative AI can lead to “surreal hallucinations or uncomfortable conversations, resulting in customers repelling the brand forever.”
While many companies have specific guidelines and parameters to fine-tune responses and hopefully avoid these situations, Mukherjee notes that these parameters will evolve and train as time goes on. Early on, however, content can result in biased and insensitive content, and it’s especially important that consumer goods companies keep an eye for tech providers without transparent data sources and usage policies.
There are several limitations, says Gartner, which notes that ChatGPT, for example, is only trained on data through 2021, cannot cite its sources, does not provide data privacy assurances, and does not currently have a supported API available.
Much of the risks are due to the vastness of generative AI, which generates text and graphics based on the huge amount of data it has processed, says Harmon. This could include personal information and copyrighted info, with the potential to generate incorrect information since there is no human involved in the process.
The best strategy at present, says Harmon, is “cautious experimentation.”