Artificial Intelligence Prompt Engineering as a New Digital Competence: Analysis of Generative AI Technologies Such as ChatGPT

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

This prompt is more abstract, potentially yielding surprising and inspirational results for further exploration.

“This prompt is more abstract, potentially yielding surprising and inspirational results for further exploration.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

This prompt offers a precise description leading to more predictable and concrete imagery.

“This prompt offers a precise description leading to more predictable and concrete imagery.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

The instruction 'Briefly and without unnecessary details' is not a clear instruction.

“'Briefly and without unnecessary details' is not a clear instruction.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The prompt example 'Tell me what prompt engineering is. Write briefly and without unnecessary details.' does not provide a clear structure for the response and does not indicate the intended audience.

“This example does not provide a clear structure for the response, nor does it indicate the intended audience.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

The prompt example 'Write a 3-5 sentence explanation of what prompt engineering is. Write the response as if it were directed to a high school student.' provides a clear response structure and specifies the intended audience, signaling the required tone and vocabulary.

“This example provides a clear response structure and specifies the intended audience and thus signals the required tone and vocabulary to use in the response.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

Prompt engineering is a creative skill for generating AI art, and expertise and practice are required to master this skill.

“Oppenlaender, Visuri, Paananen, Linder, and Silvennoinen (2023) investigated prompt engineering as a creative skill for generating AI art, emphasizing that expertise and practice are required to master this skill.”

• cites Oppenlaender, Visuri, Paananen, Linder, and Silvennoinen, 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based).

Prompt engineering allows for greater control over the output, as the right prompts can help ensure AI models create the desired content.

“It allows for greater control over the output, as the right prompts can help ensure AI models create the desired content.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

Digital competences transcend mere technical abilities and represent a complex concept that integrates a wide range of skills, including cognitive, emotional, and sociological knowledge, which are crucial for effectively navigating digital environments.

“Scholars indicate that digital competences transcend mere technical abilities and represent a complex concept that integrates a wide range of skills, including cognitive, emotional, and sociological knowledge, which are crucial for effectively navigating digital environments (Røkenes & Krumsvik, 2016;Wang, Zhang, Wang, & Li, 2021).”

• cites Røkenes & Krumsvik, 2016Not checked yet

  Not checked in the free run.

• cites Wang, Zhang, Wang, & Li, 2021Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

The construction of appropriate prompts depends on the tool used at the moment, even though there are general best practices for visual models.

“Noteworthy, the construction of appropriate prompts depends on the tool used at the moment, even though there are general best practices for visual models.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Prompt engineering concentrates on manually creating and revising prompts, aligning it more with human-centred fields such as human-computer interaction and conversational AI, rather than the general field of machine learning.

“In contrast to these approaches, prompt engineering concentrates on manually creating and revising prompts, aligning it more with human-centred fields such as human-computer interaction, and conversational AI, rather than the general field of machine learning (Oppenlaender, Linder, & Silvennoinen, 2023).”

• cites Oppenlaender, Linder, & Silvennoinen, 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

An effective prompt may include the following elements: context, instruction, input data, and expected output format.

“Ultimately, an effective prompt may include the following elements: 1) Context -can include information about the role the model should play for the task or any necessary information about the situation that may ground it ('You are a salesperson in a technology company');”

• Also stated in: § Text-to-text Prompting · plain

  “2) Instruction -the task to be performed ('Write an email to a customer thanking for high attendance during the last workshop on MLOps'); 3) Input Data -data for few-shot learning or information the model should use to perform the task ('Customer name: XYZ, keywords to include = [future, innovation, parthership]'); 4) Expected output format -information about the format and type of output in which”

Classified as probably not a substantial claim.

The DigComp framework is structured around five core competence areas: information and data literacy, communication and collaboration, digital content creation, safety, and problem-solving.

“The DigComp framework is structured around five core competence areas.Since 2016, these areas include information and data literacy, communication and collaboration, digital content creation, safety, and problem-solving (VUORIKARI Rina et al., 2022).”

• cites VUORIKARI Rina et al., 2022Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

AI prompt engineering has profound implications across various domains such as entrepreneurship, art, science, and healthcare.

“The study revealed the profound implications of AI prompt engineering across various domains such as entrepreneurship, art, science, and healthcare.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Zero-shot prompting enables a model to make predictions about previously unseen data without the need for any additional training.

“Zero-shot prompting enables a model to make predictions about previously unseen data without the need for any additional training.”

• cites Zhou, He, Ma, Berg-Kirkpatrick, & Neubig, 2022Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Building prompts for text-to-image models is conceptually simpler than for text-to-text models, as the number of tasks is limited in their case.

“Building prompts for text-to-image models is conceptually simpler than for text-to-text models, as the number of tasks is limited in their case.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing).

Digital competences consist of a blend of knowledge, skills, and attitudes encompassing concepts, facts, skill descriptions, and attitudes.

“Digital competences consist of a blend of knowledge, skills, and attitudes encompassing concepts, facts, skill descriptions (for example, the capability to execute processes), and attitudes (such as a mindset or inclination to take action) (VUORIKARI Rina, Kluzer, & Punie, 2022).”

• cites VUORIKARI Rina, Kluzer, & Punie, 2022Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based).

Effective prompts contribute to improved accuracy by guiding AI models to generate more relevant and valuable content.

“Additionally, effective prompts contribute to improved accuracy by guiding AI models to generate more relevant and valuable content.”

Classified as probably not a substantial claim.

The interactive feature of ChatGPT enables human collaboration with the underlying LLM to improve its performance through a multi-turn 'prompt engineering' fashion.

“However, the interactive feature of ChatGPT enables human collaboration with the underlying LLM to improve its performance through a multi-turn 'prompt engineering' fashion, i.e. an interactive approach in which users iteratively engage with a large language model (LLM) like ChatGPT in a conversational manner.”

Classified as probably not a substantial claim.

The DigComp framework offers a more in-depth and structured approach to understanding digital competence.

“One of the most comprehensive developments in this area is the DigComp framework which offers a more in-depth and structured approach to understanding digital competence (Reisoğlu & Çebi, 2020).”

• cites Reisoğlu & Çebi, 2020Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

AI generative technologies, such as ChatGPT or Bard, have significantly impacted the business landscape, transforming the way organizations operate, communicate, and innovate.

“Furthermore, AI generative technologies, such as ChatGPT or Bard, have significantly impacted the business landscape, transforming the way organizations operate, communicate, and innovate (Korzynski, Kozminski, & Baczynska, 2023;Wach et al., 2023).”

• cites Korzynski, Kozminski, & Baczynska, 2023Not checked yet

  Not checked in the free run.

• cites Wach et al., 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Reynolds and McDonell (2021) introduced the concept of a 'metaprompt,' which is a higher-order prompt designed to seed the language model to generate its own natural language prompts for a range of tasks.

“Reynolds and McDonell (2021) introduced the concept of a 'metaprompt,' which is a higher-order prompt designed to seed the language model to generate its own natural language prompts for a range of tasks.”

• cites Reynolds and McDonell, 2021Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

LLMs have legal and ethical risks, including stochastic parrots.

“Li (2023) highlighted the legal and ethical risks associated with LLMs, including stochastic parrots.”

• cites Li, 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based).

AI prompt engineering has emerged as a critical aspect of interacting with large language models (LLMs) such as ChatGPT.

“In recent years, AI prompt engineering has emerged as a critical aspect of interacting with large language models (LLMs) such as ChatGPT.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based). May combine two claims.

For models returning text, the creation of prompts usually relies on the desire to solve a specific language task, therefore, the prompt architecture differs from task to task.

“For models returning text, the creation of prompts usually relies on the desire to solve a specific language task, therefore, the prompt architecture differs from task to task.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Prompt engineering can enhance creativity by presenting AI models with new and unique prompts to explore.

“Furthermore, prompt engineering can enhance creativity by presenting AI models with new and unique prompts to explore (God, 2023).”

• cites God, 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

If more than 2-4 objects are placed in an image prompt, generators usually start to become less precise and create blurred content.

“5) We should not place more than 2-4 objects in the image.If the number is higher than the generators usually start to become less precise and create blurred content.”

• cites Oppenlaender, 2022Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The ChatExtract method using engineered prompts applied to a conversational LLM can accurately extract data from research articles.

“Polak and Morgan (2023) proposed the ChatExtract method using engineered prompts applied to a conversational LLM for accurate data extraction from research articles.”

• cites Polak and Morgan, 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The more specific and detailed the prompt is, the greater the chance of getting the desired result.

“6) It is important to be specific as the more specific and detailed the prompt is, the greater the chance of getting the desired result (Oppenlaender, 2022).”

• cites Oppenlaender, 2022Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

Few-shot prompting is a technique in which the model receives a small number of examples, typically between two and five, to quickly adapt to new tasks.

“Meanwhile, few-shot prompting is a technique, in which the model receives a small number of examples, typically between two and five, to quickly adapt to new tasks.”

• cites Meyerson et al., 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The metaprompt approach can potentially lead to more accurate and creative results, as the language model leverages its vast knowledge to generate suitable prompts.

“This approach can potentially lead to more accurate and creative results, as the language model leverages its vast knowledge to generate suitable prompts.”

Probably not a substantial claim: may not pass S3 (load-bearing).

White et al. (2023) presented a catalogue of prompt engineering patterns that can be applied to solve common problems in LLM conversations.

“White et al. (2023) presented a catalogue of prompt engineering patterns that can be applied to solve common problems in LLM conversations.”

• cites White et al., 2023Not checked yet

  Not checked in the free run.

May combine two claims.

This study highlights the potential for AI prompt engineering to streamline data extraction processes and improve the efficiency of research across scientific disciplines.

“This study highlights the potential for AI prompt engineering to streamline data extraction processes and improve the efficiency of research across scientific disciplines.”

Classified as probably not a substantial claim.

Prompt engineering methods are continually refined to help large language models adapt to new tasks with limited or no supervised training.

“These methods are being refined continually to help LLMs adapt to new tasks with limited or no supervised training.”

Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

This approach would yield graphics that adhere to a uniform style, thereby establishing a cohesive visual identity for the website.

“This approach would yield graphics that adhere to a uniform style, thereby establishing a cohesive visual identity for the website (Parsons, 2023).”

• cites Parsons, 2023Not checked yet

  Not checked in the free run.

May combine two claims.

Prompt engineering skill can lead to more accurate and contextually relevant results, which in turn improves the overall performance and efficiency of AI models.

“Such a skill can lead to the generation of more accurate and contextually relevant results, which in turn improves the overall performance and efficiency of the AI models.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

In art, prompt engineering allows for new forms of creative expression.

“In art, it allows for new forms of creative expression.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Ekin (2023) provides a comprehensive guide on prompt engineering techniques, tips, and best practices for ChatGPT.

“Ekin (2023) provides a comprehensive guide on prompt engineering techniques, tips, and best practices for ChatGPT, a powerful tool in natural language processing (NLP) with applications across various industries.”

• cites Ekin, 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing). May combine two claims.

This theme showcases the artistic potential of prompt engineering, opening up new avenues for creative expression and collaboration between humans and AI.

“This theme showcases the artistic potential of prompt engineering, opening up new avenues for creative expression and collaboration between humans and AI.”

Probably not a substantial claim: may not pass S4 (based).

In healthcare, prompt engineering contributes significantly to applications such as question-answering systems, text summarization, and machine translation.

“Finally, in healthcare, it contributes significantly to applications like question-answering systems, text summarization, and machine translation.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Communities such as Learn Prompting provide a thriving environment that fosters learning and growth in the AI field.

“Such communities provide a thriving environment that fosters learning and growth in the AI field.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Ling et al. (2023) explored the use of reasoning chains as a prompting technique, which leverages structured sequences of related knowledge statements from external knowledge bases to improve the model's performance in open-domain commonsense reasoning tasks.

“Ling et al. (2023) explored the use of reasoning chains as a prompting technique, which leverages structured sequences of related knowledge statements from external knowledge bases to improve the model's performance in open-domain commonsense reasoning tasks.”

• cites Ling et al., 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

There is a need for digital competences as a core, fundamental literacy addressing both technology mastery and a digital citizenship mindset, especially in higher education.

“These studies indicate the need for digital competences as a core, fundamental literacy addressing both technology mastery and a digital citizenship mindset, especially in higher education (Martzoukou, Fulton, Kostagiolas, & Lavranos, 2020).”

• cites Martzoukou, Fulton, Kostagiolas, & Lavranos, 2020Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing).

The DigComp framework acknowledges the critical role digital skills play in the constantly changing digital landscape of today.

“Created with the purpose of informing and guiding policies to enhance digital competencies for citizens of all age groups, the framework acknowledges the critical role digital skills play in the constantly changing digital landscape of today (Štaka, Vuković, & Vujović, 2022).”

• cites Štaka, Vuković, & Vujović, 2022Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Non-AI experts often struggle with prompt engineering, indicating a need for more accessible and systematic approaches.

“Furthermore, non-AI experts often struggle with prompt engineering, indicating a need for more accessible and systematic approaches.”

Classified as probably not a substantial claim.

The AI PROMPT framework provides clear, actionable recommendations for text-to-text prompt engineering, helping practitioners refine their prompts to elicit better responses from language models.

“It provides clear, actionable recommendations for text-to-text prompt engineering, thus helping practitioners refine their prompts to elicit better responses from language models.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The majority of studies on digital competences are related to the education industry, focusing on students' and teachers' abilities to use technology effectively in learning and teaching.

“The majority of studies on digital competences are related to the education industry, focusing on students' and teachers' abilities to use technology effectively in learning and teaching (Fernández-Batanero, Montenegro-Rueda, Fernández-Cerero, & García-Martínez, 2022; Pozo-Sánchez, López-Belmonte, Rodríguez-García, & López-Núñez, 2020) highlighting the need for continued improvement in this area (”

• cites Fernández-Batanero, Montenegro-Rueda, Fernández-Cerero, & García-Martínez, 2022Not checked yet

  Not checked in the free run.

• cites Pozo-Sánchez, López-Belmonte, Rodríguez-García, & López-Núñez, 2020Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S4 (based).

By following the methods, guidelines, and best practices discussed in the article, users can obtain more accurate, relevant, and fitting responses from both text-to-image and text-to-text generative models.

“By following the methods, guidelines, and best practices discussed in this article, users can acquire more accurate, relevant, and fitting responses from both text-to-image and text-to-text generative models.”

Probably not a substantial claim: may not pass S2 (checkable).

This study underscores the potential for AI prompt engineering to revolutionize content creation and communication strategies in the world of entrepreneurship.

“This study underscores the potential for AI prompt engineering to revolutionize content creation and communication strategies in the world of entrepreneurship.”

Probably not a substantial claim: may not pass S4 (based). May combine two claims.

Challenges remain in prompt engineering, including hallucination problems and the tendency of large language models to mimic and reproduce existing text without deep understanding of the content.

“However, there are still challenges to be addressed, including hallucination problems and the tendency for LLMs to mimic and reproduce the existing text without a deep understanding of the content.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

In science, prompt engineering can streamline data extraction processes from research articles.

“In science, it can streamline data extraction processes from research articles.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The metaprompt approach aims to leverage the language model's inherent capabilities and understanding of natural language to create more effective and nuanced prompts.

“The metaprompt approach aims to leverage the language model's inherent capabilities and understanding of natural language to create more effective and nuanced prompts.”

• cites Reynolds and McDonell, 2021Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S2 (checkable).

Prompt engineers will hold an essential position in the realm of artificial intelligence.

“Although this statement might be somewhat overstated, prompt engineers will undoubtedly hold an essential position in the realm of artificial intelligence.”

Probably not a substantial claim: may not pass S3 (load-bearing).

The literature focuses strongly on developing prompt engineering techniques such as metaprompts, reasoning chains, and zero-shot and few-shot prompting.

“The literature reveals a strong focus on the development of prompt engineering techniques, such as the use of metaprompts and reasoning chains, as well as zero-shot and few-shot prompting techniques.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Recently, there has been a surge in the number of text-to-image generation models based on LLM (DALL•E2, Stable Diffusion, Midjourney).

“Recently, there has been a surge in the number of text-to-image generation models based on LLM (DALL•E2, Stable Diffusion, Midjourney).”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

Each task is individual and needs a specific matching prompt, so the presented examples should be treated only as a starting point for tests.

“Remember to treat these examples as a good starting point for tests, as each task is individual and needs a specific matching prompt.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based).

Communities play a pivotal role in the AI and prompt engineering landscape.

“Communities also play a pivotal role in the AI and prompt engineering landscape.”

Classified as probably not a substantial claim. May combine two claims.

By carefully formulating prompts, organizations can derive increased value from their AI systems, because the models will then deliver more precise and helpful outputs.

“By carefully formulating prompts, these entities can derive increased value from their AI systems, as the models will then deliver more precise and helpful outputs.”

Probably not a substantial claim: may not pass S2 (checkable), S3 (load-bearing), S4 (based). May combine two claims.

In entrepreneurship, prompt engineering can mimic celebrity CEO archetypes, revolutionizing content creation and communication strategies.

“In entrepreneurship, it can mimic celebrity CEO archetypes revolutionizing content creation and communication strategies.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

For conversational models such as ChatGPT, feedback on the model's responses can help it better understand the desired output.

“For conversational models, such as ChatGPT, feedback on the model's responses can help it better understand the desired output.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Scholars emphasize the importance of critically and reflectively employing technology to generate new insights and knowledge.

“Moreover, scholars emphasize the importance of critically and reflectively employing technology to generate new insights and knowledge (DeWitt & Alias, 2023;Głodowska, Maciejewski, & Wach, 2023).”

• cites DeWitt & Alias, 2023Not checked yet

  Not checked in the free run.

• cites Głodowska, Maciejewski, & Wach, 2023Not checked yet

  Not checked in the free run.

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

Providing a model with a few examples of expected responses (few-shot learning) can help it adapt its style and way of responding.

“Moreover, providing the model with a few examples of expected responses (few-shot learning) can help it adapt its style and way of responding.”

Probably not a substantial claim: may not pass S3 (load-bearing).

Open-source frameworks such as OpenPrompt provide an environment for users to learn, experiment, and design their own prompts, fostering the development of prompt engineering skills.

“They provide a conducive environment for users to learn, experiment, and even design their own prompts, fostering the development of skills in prompt engineering.”

Probably not a substantial claim: may not pass S4 (based).

Prompt engineering has shown promising results in a variety of domains.

“Despite these challenges, prompt engineering has shown promising results in a variety of domains.”

Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

In many tasks, several prompt building blocks need to be combined to receive the expected result.

“In many tasks, we need to combine the following building blocks to receive the expected result.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

The DigComp framework, with its five core competence areas (information and data literacy, communication and collaboration, digital content creation, safety, and problem-solving), provides a valuable lens to understand and improve AI prompt engineering competencies.

“The DigComp framework, which outlines five core competence areas, i.e. information and data literacy, communication and collaboration, digital content creation, safety, and problem-solving provides a valuable lens to understand and improve AI prompt engineering competencies.”

Classified as probably not a substantial claim.

Hallucination and mimicry problems can lead to the generation of false, misleading, biased, or inappropriate information.

“These issues can lead to the generation of false, misleading, biased, or inappropriate information.”

Probably not a substantial claim: may not pass S2 (checkable), S4 (based).

Implementing efficient prompting strategies is essential for improving the performance of large language models in all possible applications.

“The implementation of efficient prompting strategies is essential for improving the performance of LLMs in all possible applications.”

Probably not a substantial claim: may not pass S3 (load-bearing).

A wide variety of frameworks and models have been developed and utilized to explore and understand digital competence from different perspectives.

“Over the years, a wide variety of frameworks and models have been developed and utilized to explore and understand digital competence from different perspectives (Blundell, Mukherjee, & Nykvist, 2022;Guitert, Romeu, & Baztán, 2021;Kelentrić, Helland, & Arstorp, 2017).”

• cites Blundell, Mukherjee, & Nykvist, 2022Not checked yet

  Not checked in the free run.

• cites Guitert, Romeu, & Baztán, 2021Not checked yet

  Not checked in the free run.

• cites Kelentrić, Helland, & Arstorp, 2017Not checked yet

  Not checked in the free run.