By Prof. Dr. Julia Glesner (University of Applied Sciences Potsdam) and Sara Tvrdišić (University of Arts in Belgrade)
/TEACHING EXPERIENCES
How to use ChatGPT in Cultural Marketing?
Introduction
In November 2022, the US company OpenAI launched its chatbot ChatGPT built on top of its large language models (LLMs). “This chatbot has redefined the standards of artificial intelligence, proving that machines can indeed ‘learn’ the complexities of human language and interaction”, urteilte Forbes in May 2023 (Marr 2023).
As new trends constantly emerge worldwide, such as ChatGPT, it is necessary to approach the promotion of these tools to the broader population in an appropriate manner. New technologies and artificial intelligence tools simultaneously offer enhanced functionality but also the potential for misuse, such as plagiarism. Therefore, in the home countries of the authors, various discussions have been initiated at numerous universities and other organisations over the years regarding the advantages and disadvantages of these tools and how to best utilize them. It will be important for university teachers to actively integrate tools such as ChatGPT into their teaching and to critically reflect on the necessary competences for the use of such a chatbot. To achieve such learning objectives, the first step is to understand how ChatGPT works and how it can be used. In the following, the focus is on one such application, namely how ChatGPT can be used in cultural marketing.
Definition of Artificial Intelligence:
Artificial intelligence (AI) refers to programmes and applications based on algorithms
that aim to emulate human intelligence as closely as possible through machine learning processes. The European Parliament offers a concise, descriptive definition:
“AI is the ability of a machine to display human-like capabilities such as reasoning, learning, planning and creativity. AI enables technical systems to perceive their environment, deal with what they perceive, solve problems and act to achieve a specific goal. The computer receives data – already prepared or gathered through its own sensors such as a camera – processes it and responds. AI systems are capable of adapting their behaviour to a certain degree by analysing the effects of previous actions and working autonomously.“ (European Parliament 2023)
How does ChatGPT work?
A chatbot such as ChatGPT is a software application or web interface that aims to mimic human conversation through text or voice interactions. Modern chatbots are typically online and use AI systems that can maintain a conversation with a user and simulating the way a human would behave as a conversational partner. Besides ChatGPT, there are a number of other chatbots, e.g. Bing Chat, Google Bard, Einstein GPT, Deep Converse and others.
ChatGPT has been developed by the US-American company OpenAI. GPT stands for Generative Pre-trained Transformer. It is a so-called conversational AI system in contrast to e.g. rule-based chatbots. Rule-based chatbots use keywords and other language identifiers to trigger pre-written responses (Bishop 2023).
“In comparison, AI chatbots that use machine learning understand the context and intent of a question before formulating a response. These chatbots generate their own answers to more complicated questions using natural-language responses.” (Alburger 2023)
Machine learning is a field of artificial intelligence where computers learn from data to improve their performance on a specific task without being explicitly programmed, by identifying patterns and making predictions or decisions based on that data.
AI language models such as ChatGPT are neural networks that have been trained to generate texts. A neural network is a computational model inspired by the human brain’s structure, consisting of interconnected nodes (neurons) organized in layers. It’s used in machine learning to process and learn from data, with input, hidden, and output layers, allowing it to make complex decisions or predictions through weighted connections between neurons. These neural networks are trained through extensive training with large amounts of text corpora to determine statistical probabilities for the occurrence of words within certain contexts (for further explanation see Vítek 2023).
ChatGPT is a text-based chatbot that can interact with humans and is trained to produce as natural speech as possible. Based on prompts, i.e., requests and questions to the AI, ChatGPT generates answers and takes the context of the chat history into account (Thompson 2022).
This example uses the prompt Act as a digital marketer in the cultural sector. How can ChatGPT used for cultural marketing? Answer shortly. and shows the corresponding result CHatGPT gave.At the latest since the ChatGPT service was released for individual use in November 2022, not only the academic world has been in an uproar. It generated worldwide interest in its diverse applications for personal and professional use, including customer service. From the wide field of cultural marketing, we have selected four opportunities for you where we believe ChatGPT can be profitably used. These are optimization for search engines (SEO), client communication, semi-automatic creation of advertising texts and the automatisation of social media communication.
SEO optimization:
Using ChatGPT or similar AI models for SEO[1] can be a creative and effective approach to improve website’s content and ranking, which can be used in a cultural marketing too.
ChatGPT can conduct keyword research, which means to start by conducting keyword research to identify relevant keywords and phrases for some content. Furthermore, it can be used to generate keyword ideas. ChatGPT can also help generate high-quality content around target keywords, input keyword or topic, and can provide with content ideas, outlines, or even full paragraphs, including using keywords naturally, having a clear structure, and providing valuable information to the reader. ChatGPT can utilize to help craft compelling meta titles and descriptions that include target keywords. These elements play a crucial role in click-through rates from search engine results pages. It can also to expand existing content for example of blog post or article and can suggest additional sections or points to make content more comprehensive and valuable to readers. ChatGPT can create FAQ[2] sections on websites, which can improve chances of appearing in featured snippets on Google. It can suggest improvements, such as adding headers, bullet points, or multimedia elements like images and videos. ChatGPT can localize content by translating and adapting it for different languages and regions, additionally assisting in generating ideas for content calendar, suggesting trending topics, seasonal content, or evergreen content ideas. ChatGPT can provide with regular updates on emerging trends and topics, which can help to stay ahead in content strategy, and encourage user engagement to generate interactive content like quizzes, polls, or surveys that can attract and retain website visitors.
Tough ChatGPT can be a valuable tool for content generation and optimization, it should only complement your human expertise and understanding of your target audience.
Client Communication:
Even though, according to a survey, every fourth user in Germany can imagine communicating with a chatbot (Statista, 2023), communication robots are not yet accepted in all parts of the population. Many people are skeptical, especially about data protection and privacy, when it comes to digital assistants as a further development of chatbots.
While the technology has impressive writing and parsing capabilities, you need a programmer to customise the model available in the API and develop customer service features that support your organisation’s unique needs.
Some internationally known companies have already adapted ChatGPT for their needs via an app. Collaborative workspace platform Slack has created an app allowing its users to use ChatGPT for managing workflows and communicating with colleagues (Slack 2023). Users of the plugin always have an assistant on hand to answer questions and offer suggestions on the best way to move forward with the projects they are working on.
Semi-Automatic Creation of Advertising Texts:
The semi-automatic creation of advertising texts using ChatGPT refers to a process where AI, like ChatGPT, assists human copywriters in generating compelling advertising content. This approach combines the creativity and strategic thinking of humans with the speed and assistance of AI to streamline the advertising copywriting process.
Usually firstly, human marketers identify the key topics, products, or services they want to promote and the relevant keywords. Than ChatGPT is used to provide content suggestions based on the selected keywords and topics. The AI can generate headlines, product descriptions, and even entire advertising copies. Secondly, AI-generated content is reviewed and customized by human copywriters. In this phase, they can add branding elements, emotional appeal, and ensure the content aligns with the cultural project’s voice and style. The advertising copies, both AI-generated and human-edited, can be tested using A/B testing to measure which versions perform better with the target audience. This data-driven approach helps refine the content further. Finaly, marketers can use AI to analyze the performance data and make data-driven decisions on how to optimize advertising copies for better results.
The advantages of this approach include (time) efficiency, variety and data-driven decision-making. Those means that AI can assist in quickly generating initial drafts of advertising content, saving time for human copywriters, suggest diverse angles and ideas for advertising campaigns, and marketers can use it to analyze data and optimize advertising copies for better return of investment.
Automatisation of Social Media Communication:
Anyone who has worked in social media marketing knows how many hours go into planning, copywriting and posting content. Social media marketing is a time-consuming task. Its potential lies in being able to work in a more target group-specific way. ChatGPT offers the possibility simpler posts to be written semi automatically and to schedule posts to be sent at a specific time. This can make social media marketing more time efficient.
The Irish digital marketing agency Glassfull Media offers a film on the basic workflow based on Google sheets and docs (Glassfull Media 2023). Basically, the workflow Glassfull Media presents is based on connecting Google Excel, ChatGPT and a dashboard. The prerequisite for this workflow is that, firstly, you work with a dashboard for organising your social media posts and, secondly, you are willing to use Google and, thirdly, to pay a fee for this.
The example of Glassfull Media shows how precisely the prompts must be formulated:
“Act as a digital marketer. Create social media posts for these annual dates. In a tone of educational or funny. Include 2 hashtags and icons. Target audience of Ireland small business.”
The content the company uses – Irish festivity dates – might not be typical for cultural organization. For a theatre company f.ex., you will have to pre-produce stage photographs or films and write texts accordingly to the stage concept etc. The workflow will not be as fast for all use cases in the cultural sector, but it offers the potential for significant time savings.
Questions for further discussions:
There are no official figures on how much energy the training of ChatGPT consumed. Estimates put it at 1,287 megawatt hours, which is roughly equivalent to the annual energy consumption of 120 US households. It is true that Microsoft, in whose cloud ChatGPT runs, operates its data centers partly with renewable energies. But given the immense computing power, the bot is unlikely to be climate neutral.
A single request of about 230 words consumes 581 watt-hours. The one billion requests to Chat-GPT in February 2023 would thus have caused a consumption of 581 gigawatt hours, roughly the electricity consumption of a city with 170,000 inhabitants per year. That corresponds to around 244,000 tonnes of carbon dioxide. This means that the potential that ChatGPT offers in increasing efficiency is set against our responsibility to act in a climate neutral way – a trade-off that we must make very carefully.
References:
Alburger, J. (2023) Rule-Based Chatbots vs. AI Chatbots: Key Differences. https://www.hubtype.com/blog/rule-based-chatbots-vs-ai-chatbots
Бейкер, П. (2023) ChatGPT для чайников. Hoboken: John Wiley & Sons, Inc.
Bishop, C. (2023) Chatbots vs. conversational AI: What’s the difference? https://www.zendesk.de/blog/chatbot-vs-conversational-ai/#:~:text=Chatbots%20are%20a%20type%20of,built%20on%20conversational%20AI%20technology.
European Parliament (2023) What is artificial intelligence and how is it used?
Glassfull Media (2023) How to use ChatGPT to create social media posts (super fast). Yoututbe film. https://www.youtube.com/watch?v=QJaGqIWNpyE
Ibrahim, J. (2023) The Art of Asking ChatGPT for High-Quality Answers: A Complete Guide to Prompt Engineering Techniques. Nzunda Technologies Limited
Marr, B. (2023) A Short History Of ChatGPT: How We Got To Where We Are Today
Forbes, May 19, 2023. https://www.forbes.com/sites/bernardmarr/2023/05/19/a-short-history-of-chatgpt-how-we-got-to-where-we-are-today/
Sarrion, E. (2023) Exploring the Power of ChatGPT: Applications, Techniques, and Implications. Paris: Apress.
Slack (2023) Why we built the ChatGPT app for Slack. A conversation with Simón Posada Fishman, the OpenAI solutions engineer who helped develop the breakthrough app.
https://slack.com/intl/en-gb/blog/news/why-we-built-the-chatgpt-app-for-slack
statista.de (2023) Wie oft benutzen Sie ChatGPT im Durchschnitt? (How often do you use ChatGPT on average?) https://de.statista.com/statistik/daten/studie/1401298/umfrage/durchschnittliche-nutzung-chatgpt/#:~:text=Rund%20neun%20Prozent%20der%20Befragten,Befragten%2C%20die%20ChatGPT%20nie%20verwenden.
Thompson, A. (2022) The ChatGPT Prompt Book https://lifearchitect.ai/chatgpt-prompt-book/
Vítek, F. (2023) How ChatGPT really works in simple words (and pictures). https://mocnedata.sk/en/how-chatgpt-works-in-simple-words/
Wolfram, S. (2023) What Is ChatGPT Doing … and Why Does It Work? Champaign: Wolfram Research.
Biography:
Julia Glesner is Professor of Cultural Management at Potsdam University of Applied Sciences. Her mainresearch interests are theatre management, cultural policy, cultural governance. She formerly held management positions at the Klassik Stiftung Weimar (2006-2018) and at the Erfurt Opera (2003-2006). She recently published “Opera for all. The Biography of Sir Peter Jonas”, the former state director of the Bavarian State Opera Munich, at Suhrkamp/Insel.
Sara Tvrdišić (b. 1996) is a PhD candidate at the Faculty of Dramatic Arts in Belgrade, with a scholarship for scientific research work from the Ministry of Science and Technological Development and Innovation of the Republic of Serbia. In parallel with her academic studies, she obtained a degree from the Information Technology School and defended the final exam thesis Application of biological and artificial intelligence in electronic commerce. From 2015 to 2022, she was a member of the editorial team of Mensa’s magazine MozaIQ and, as of 2022, is Mensa’s coordinator for Belgrade (Mensa is the high IQ society, for individuals who have scored in the top 2% on a recognized intelligence test). She won numerous international awards in the field of art photography, as well as national and city awards in many different fields (mathematics, chess, etc.).
[1] Search Engine Optimization
[2] Frequently Asked Questions