Concerns are rapidly growing over the capabilities of OpenAI’s chatbot, ChatGPT, which can answer questions in an alarmingly quick fashion with a very respectable degree of accuracy.
The chatbot marks probably the biggest breakthrough in artificial intelligence (AI) with regards to everyday applications, and in doing so, poses the largest threat to the job market so far.
Since its launch as a prototype in November 2022, ChatGPT quickly gained attention from academics for its ability to write copy at speed in a conversational manner, threatening the authenticity of the education sector.
Following the release of ChatGPT’s beta, OpenAI was valued at £23.4 billion ($29bn). While OpenAI, which is currently in discussions with Microsoft over potential investment, said that its chatbot “sometimes writes plausible-sounding but incorrect or nonsensical answers,” it has drawn concern from many for its potential ability to eradicate jobs in certain sectors.
But how could the latest AI innovation impact the retail technology sector and what are the implications of it doing so?
The capabilities of ChatGPT
ChatGPT is essentially a large language model that is able to formulate written copy as if it was human-written. It’s trained on a large dataset of written works from the internet. This is why those working in the education sector have become wary of how it will impact the work of students.
The AI tool is able to write essays, emails, poetry, answer questions, or generate lines of code based on just a small human-inserted prompt.
According to AI-powered content creation platform, Jasper, about 80,000 clients have used its software to draft ads, emails, blogs and other material.
“There are so many applications, it’s impossible to even start to list them,” Publicis Sapient executive vice president of retail Guy Elliot tells Charged.
“Acceleration and automation of the definition of navigation and taxonomy in ecommerce as well as dynamic creation of product descriptions, for example. This would allow dynamic loading and automatic categorisation of product catalogues, removing the massive amount of manual labor currently needed,” he notes.
It would also allow instant support for many languages through high quality automatic translation (not necessarily ChatGPT currently, but AI technology like Deep), making products more attractive to a larger audience.
Elliot also cited the wealth of data and customer insight that AI bots such as ChatGPT could bring to the table.
“Deep data analysis on purchasing patterns, customer reviews, sentiment analysis, promotion success and returns data in near real time to give retailers incredibly strong insights into their business, product development. This would accelerate everything from the product lifecyle, to purchasing and through inventory management and routing,” he says.
For SMEs, it could be a game changer. Elliot says that it could extradite the entire process of outsourcing the construction of an ecommerce shop.
In turn, this could also have a negative impact on proprietary ecommerce platforms such as Shopify.
“Automation of the creation of entire e-commerce shops with no need for manual programming, allowing retailers to spin up and shut down category, subsegment or even targeted ‘individual’ stores very quickly and cheaply,” he explains.
In order to test the functionality of ChatGPT, Charged assumed the role of a retail technology software firm trying to advertise its product.
We asked it a simple command: “Write a product description for a retail technology software company”
It responded in about 10 seconds with a paragraph that read:
“Introducing our retail technology software company, designed to revolutionise the way you manage and grow your business.
“Our software offers a wide range of features to streamline your operations, including inventory management, sales tracking, customer relationship management, and much more.
“With real-time data and analytics, you’ll always have a clear view of your business performance and can make informed decisions to drive growth.
“Additionally, our software is user-friendly and customizable, so you can tailor it to the unique needs of your business. Upgrade to our retail technology software and experience the benefits of streamlined operations and increased efficiency. Contact us today to learn more.”
Next we asked it to write a product description for a wool blend sweater:
“Introducing our cozy and stylish sweater, perfect for keeping warm during the colder months.
“Made with a soft, wool blend material, this sweater is comfortable and durable. The classic design features a crew neckline and long sleeves, making it versatile and easy to pair with any outfit. The ribbed cuffs and hem add a touch of texture and visual interest.
“Available in a variety of colors, this sweater is a must-have for any wardrobe. Keep warm in style with our wool blend sweater.”
It’s clear that ChatGPT can fulfill basic requests to a very sufficient level, potentially removing a lot of the legwork needed to scale up an ecommerce business.
Product descriptions, site copy, employee notice emails, HR notifications and – perhaps most importantly – customer service jobs could all be dealt with almost exclusively by a model like ChatGPT, perhaps putting an end to customers taking umbrage with the robotic nature of current customer assistant chatbots.
According to technology firm Digiteum, existing AI chatbots can increase conversions and reduce cart abandonment by 30%. A more sentient bot such as ChatGPT could raise that figure exponentially.
Another area in which ChatGPT could save both time and money is the hiring process. When Charged asked the model if it could write a letter to a potential new employee, it wrote:
Dear [Employee Name],
We are pleased to offer you the role of retail worker at our company. We believe that your skills and experience make you an excellent fit for our team, and we are excited to have you on board.
As a retail worker, you will be responsible for providing excellent customer service, maintaining store displays, and assisting with store operations. We expect you to be able to work well in a team, and be able to multitask and adapt to the fast-paced environment of retail.
We would like to invite you to attend an orientation session on [Date] at [Time] in our store located at [Address]. During this session, you will have the opportunity to meet your fellow team members and learn more about the company and your role.
Please let us know if you have any questions or concerns about the role or the orientation session, and we will be happy to assist you. We look forward to welcoming you to our team.
Retail Manager [Company Name]
Chat GPT limitations
OpenAI has been quite vocal about the limitations of its new model, which is still in testing phase.
On its website, one of the model’s limitations says: “ChatGPT is sensitive to tweaks to the input phrasing or attempting the same prompt multiple times.
“For example, given one phrasing of a question, the model can claim to not know the answer, but given a slight rephrase, can answer correctly.”
One of its main drawbacks is also its primary strength; all information written by the model is taken from the internet and interpreted using its machine-learned algorithm.
At face value, ChatGPT is quite scary, however when looked at in further detail, for the mundane administrative tasks that a retailer might need to complete, it offers a convincing and cheap solution. For more creative needs however, ChatGPT still won’t be able to compete with a human counterpart.
Will it affect the job market?
According to a report from the World Economic Forum’s Future of Jobs Report, which was produced in 2020, that there is a potential for 85 million jobs to be displaced worldwide as a result of AI and robotics by 2025.
However, on the other side of the coin, a further 97 million jobs may emerge from the changes.
For an SME, ChatGPT and future language models could become integral and cheap parts of a business.
“Automated customer services and assistance may be over the horizon, though a narrower vocabulary and subject matter set is the next challenge as AI chatbots work in a specific sectors such as retail – or even a more specific space like luxury,” Publicis Sapient telecom, media and technology head Raj Shah told Charged in June last year.
ChatGPT uses a deep learning technique called “transformer architecture” to sift through several terabytes of data that contain billions of words to create answers to prompts or questions, meaning this horizon could be a lot closer than Shah could have predicted last year.
Productsup chief innovation officer Marcel Hollerbach believes that workers don’t need to be worried about AI affecting the job market massively.
“ChatGPT is going to help companies work more efficiently, which many workers will welcome, as the technology frees them up from repetitive admin tasks. It is important to understand that tools like ChatGPT are not working at 100% accuracy, meaning all outputs still need to be fact checked, revised, and edited.
“The worker with ChatGPT definitely can outperform the one without. Still, AI will only ever be as good as the data you feed into it, or in this case, the questions you ask.”
The future of AI will undoubtedly change the face of many industries, with ChatGPT probably marking the biggest development in accessible AI so far. But the general consensus on OpenAI’s new model says that it shouldn’t concern too many of us – yet.