Beyond the Query

A Meandering in the Mind(s) of ChatGPT (Part III of III)

In the first two articles of this three part series, the author introduced a brief history of her own word processing software use and preferences, then used several models within an AI tool, ChatGPT, to generate an article with a single prompt. There are some issues with prompts. For instance, the author found that the tools sometimes interpreted them as an instruction to avoid the use of subtitles, other times as an instruction to exclude subtitles from the word count. The author concludes with her findings regarding the current limits on the use of AI tools.

Read Part I and Part II

What You Missed in Part I

I introduced a brief history of my own word processing software use and preferences, then used several models within an AI tool, ChatGPT, to generate an article with a single prompt: “The article must be between 4750 and 5250 words, not including subtitles. The article must include at least seven cited sources from peer-reviewed journals only. The article must be formatted in the five-paragraph essay format. The article must be a comparison of word processing software between 1998 and 2004.” There are some issues with this prompt. For instance, am I instructing the tool to avoid the use of subtitles or am I instructing the tool to exclude subtitles from the word count? Who knows? The tool ignored the word count for all eight attempts and included subtitles in each article.

What You Missed in Part II

I discussed the choice of using the style of well-known authors or other creators and used the AI tool to generate the same article in the style of John Locke and in the style of Alfred Hitchcock. The Hitchcock style article is far and away my favorite because it contains the most entertaining language.

What is Going on Now: The Current State of the Art

In this Part III of the series, I explore the use of pre-trained AI models. The use of these models is effectively standing on others’ shoulders. There are hundreds of models built by others that are shared within the ChatGPT software alone such as one called “SQL Expert” for optimizing database queries; another, “ScholarGPT”, that has been trained on public sources and critical reading skills; and “RizzGPT” trained to help you create better messages for use on dating apps. I used ScholarGPT and Rizz GPT for Attempts 7 and 8, which are included here.

When you choose to use a pre-trained model, you must also determine your level of disclosure for the use of that model. Your determination may be governed by professional standards and codes of ethics attached to various credentials you may hold. The general disclosure of the use of an AI tool has been the subject of many an opinion piece in our industry but just how much disclosure is rarely discussed. “I used an AI tool to assist in this report,” is vastly different from “I used an AI tool pre-trained to generate dating app messages to assist in this report, your Honor.”

What is a Pre-Trained AI Model?

The left sidebar of the AI tool contains a list of each chat interaction in which I engaged, which can be trained separately and can learn any style so that each time an answer is provided, it is in that style. Think of each interaction as a trainee who must learn particular preferences and become an expert on particular topics. Each chat model in your account can be limited to specific topics, styles, and other preferences such that each answer it provides meets (or exceeds or spits on) your requirements and, importantly, learns to refine how well it meets those requirements over time. The AI tool can be trained and must be trained to use it effectively.

If you want to generate an article in your own writing style, you can upload files of your past articles so that it learns from those. You can specify which sources to use to generate information such as peer-reviewed journals or some other public sets of data sources. Doing so would be referred to as “training your AI model” or “creating an AI model” or “making a friend”.[1] A pre-trained model has had, as it sounds, training. Different types of input were provided to the AI model and its responses may have been refined with additional interaction. A person, people, or other computer interacted with the AI model to teach it what an appropriate response should be. It may be provided to the public so that additional responses can also be refined, or it may be contained within an organization so that its responses are refined in a more controlled environment.

If you were ever a fan of, choose-your-own-adventure books, there is a pre-trained model for that as well, “DeepGame”, with a brief interaction shown below.

Initial Interaction with Text-Based Storytelling Model

Conclusory Notes

Rather than treating each AI model as a question-and-answer session only, or even as a very specific question and answer session as this article series chronicles, prepare your interactions in the style of a conversation between teacher and student.[2] AI is not a simple query tool; interactions and responses can be intricately refined over time. This capability is critical to consider when demanding the inputs a person used to generate results—the inputs matter as do the training data sets and each user’s ongoing interactions. Content generation using an AI tool can yield massive efficiency gains when trained properly or when using a pre-trained model tweaked with personal preferences, buffered by the need (or personal bias) to manually review any citations and to add semicolons (see Part I).

This series details a largely singular prompt with minor variations along with each resulting article. My favorite result is the Hitchcock style article, which says more about my reading preferences than it does about anything else. It is important to note that only RizzGPT, designed for dating app messages, mentioned Lotus Word Pro and the cuddly Microsoft assistant, Clippy.[3] None of the AI models mentioned Apple’s Pages software.

I can only conclude that dating nowadays requires much more interesting content than regular articles and one can expect to be fact-checked when communicating. This conclusion is supported by the use of actual websites in RizzGPT’s sources (although the broad citations make it difficult to determine if the citation is appropriate), supported by the more flowery language in the RizzGPT response as compared to Attempts 1 and 2, and supported by the Clippy mention.

Writing Prompt on Model Trained on Scholarly Articles

Attempt 7: Pre-Trained on Scholarly Articles and Critical Writing Skills

7^th Article Attempt, Version 4 ScholarGPT, Word Count: 968

Title: Evolution of Word Processing Software: A Comparison between 1998 and 2004

Introduction

Word processing software has undergone significant evolution over the years, playing a crucial role in enhancing productivity and efficiency in document creation and management. The period between 1998 and 2004 marked a transformative era for word processing software, characterized by substantial advancements in functionality, user interface, and interoperability. This article aims to compare word processing software during this timeframe, focusing on key developments and innovations. By examining peer-reviewed sources, this analysis highlights the critical changes that defined word processing tools from 1998 to 2004.

1. Advancements in User Interface and Usability

During the late 1990s, word processing software began to shift towards more user-friendly interfaces. In 1998, popular word processors like Microsoft Word 97 and Corel WordPerfect 8 were dominant. These programs featured relatively simple interfaces, but with a learning curve for new users. The focus was on providing basic word processing functions with limited customization options.

By 2004, the user interface of word processing software had significantly improved. Microsoft Word 2003, for instance, introduced a more streamlined and intuitive interface, featuring task panes, smart tags, and enhanced toolbars (Karat, Campbell, and Fiegel, 1999). These enhancements made it easier for users to access and utilize various functions, reducing the time required to perform common tasks. The integration of wizards and templates further simplified the process of creating complex documents, catering to both novice and experienced users.

2. Enhancements in Collaboration and Interoperability

The period from 1998 to 2004 also saw notable improvements in collaboration and interoperability features of word processing software. In 1998, collaborative functionalities were rudimentary, often limited to basic track changes and comments. Real-time collaboration was largely non-existent, making it challenging for multiple users to work on a document simultaneously.

By 2004, word processing software had made significant strides in this area. Microsoft Word 2003, for example, introduced better integration with e-mail systems, allowing users to send and receive documents more efficiently (Grudin, 2002). Additionally, the ability to save documents in various formats, including PDF, enhanced compatibility across different platforms and devices. This period also marked the beginning of more sophisticated version control and document management systems, facilitating smoother collaboration and reducing the risk of data loss or duplication.

3. Development of Advanced Formatting and Editing Tools

Another critical aspect of word processing software evolution between 1998 and 2004 was the development of advanced formatting and editing tools. In the late 1990s, word processors provided basic text formatting options, such as font selection, paragraph alignment, and simple tables. However, these tools were often cumbersome and limited in scope.

By 2004, word processing software offered a broader range of formatting and editing capabilities. Microsoft Word 2003, for instance, introduced features like styles and themes, which allowed users to apply consistent formatting across entire documents with ease (Glynn, 2004). The introduction of advanced table and chart tools enabled users to create more visually appealing and informative documents. Furthermore, the integration of grammar and spell checkers improved the overall quality of written content, helping users produce error-free documents more efficiently.

4. Integration of Multimedia and Internet Capabilities

The late 1990s saw the beginning of multimedia and internet integration in word processing software. However, these features were relatively limited in scope. Users could insert images and hyperlinks, but the process was often cumbersome and lacked sophistication.

By 2004, word processing software had significantly enhanced its multimedia and internet capabilities. Microsoft Word 2003, for example, allowed users to insert a wider range of multimedia elements, including audio and video files (Wright and Reese, 2003). The ability to create and edit hyperlinks became more intuitive, enabling users to easily integrate web content into their documents. Additionally, the growing availability of online templates and resources provided users with more options for creating professional-quality documents.

5. Improvements in Security and File Management

Security and file management were also areas of significant improvement in word processing software between 1998 and 2004. In the late 1990s, document security was relatively basic, with limited options for protecting sensitive information. Users relied on simple password protection and basic encryption methods, which were often insufficient for safeguarding critical data.

By 2004, word processing software had introduced more robust security features. Microsoft Word 2003, for instance, included options for digital signatures, allowing users to verify the authenticity of documents (Adams, 2004). The integration of more advanced encryption methods provided better protection for sensitive information. Additionally, improved file management systems made it easier for users to organize and retrieve their documents, reducing the risk of data loss and ensuring better overall document management.

Conclusion

The evolution of word processing software from 1998 to 2004 marked a period of significant advancements in functionality, user interface, collaboration, formatting, multimedia integration, and security. These improvements transformed word processing tools from basic text editors into comprehensive document creation and management systems. The enhancements in usability, collaboration, and security provided users with more efficient and reliable tools for their word processing needs. As technology continued to advance, the foundations laid during this transformative period set the stage for further innovations in word processing software, shaping the tools we use today.

References

1. Adams, A. (2004). Digital signatures and word processing software: Enhancing document security. Journal of Information Security, 15(3), 45–59.
2. Glynn, M. (2004). Advanced formatting tools in Microsoft Word 2003: Enhancing document presentation. Journal of Computing and Information Technology, 12(2), 101–110.
3. Grudin, J. (2002). Collaboration and interoperability in word processing software: A comparative study. Journal of Human-Computer Interaction, 14(4), 337–359.
4. Karat, J., Campbell, R., and Fiegel, R. (1999). Interface design improvements in word processing software: A case study of Microsoft Word. International Journal of Human-Computer Studies, 50(3), 269–299.
5. Wright, P., and Reese, C. (2003). Multimedia integration in word processing software: Enhancements and challenges. Journal of Multimedia Systems, 10(4), 217–226.

Writing Prompt on a Model Trained for Dating App Messages

Attempt 8: Pre-Trained for Dating App Messages

8^th Article Attempt, Version 4 RizzGPT, Word Count: 927

A Comparison of Word Processing Software: 1998 vs. 2004

In the ever-evolving landscape of technology, the period between 1998 and 2004 witnessed significant advancements in word processing software. This transformation can be attributed to several factors, including the rise of the internet, increased competition among software developers, and the growing demands of users. This essay delves into a comparative analysis of word processing software during these years, focusing on features, user interfaces, accessibility, and overall performance. The evolution from rudimentary word processors to more sophisticated and user-friendly applications encapsulates the broader trends in technological innovation during this era.

1. Features and Functionalities: 1998 vs. 2004

In 1998, word processing software was primarily dominated by Microsoft Word 97, Corel WordPerfect 8, and Lotus Word Pro. These programs were robust for their time, offering basic formatting options, spell check, and simple graphical elements. Microsoft Word 97, for example, introduced the Office Assistant, an interactive help tool that, despite its mixed reception, represented an early attempt to make software more user-friendly (Microsoft, 1998). Corel WordPerfect 8 was known for its Reveal Codes feature, allowing users to see formatting codes, which was a boon for precise document control (Corel Corporation, 1998). Lotus Word Pro, though less popular, offered SmartMasters for creating various types of documents, showcasing an early form of templates (IBM, 1998).

By 2004, word processing software had evolved significantly. Microsoft Word 2003 emerged as the market leader, boasting advanced features like Smart Tags and an enhanced Reviewing toolbar, which facilitated collaborative editing and tracking changes (Microsoft, 2004). OpenOffice.org Writer, an open-source alternative, began gaining traction with features like robust PDF export options and support for a variety of file formats (Sun Microsystems, 2004). Additionally, Corel WordPerfect 12 introduced a RealTime Preview feature, allowing users to see formatting changes in real-time, enhancing the user experience (Corel Corporation, 2004).

2. User Interfaces: Evolution Towards Intuitiveness

The user interfaces of word processing software in 1998 were functional but often cluttered and not particularly intuitive. Microsoft Word 97, with its plethora of toolbars and the much-maligned Clippy, the Office Assistant, exemplified this cluttered approach. Corel WordPerfect 8 and Lotus Word Pro also featured dense menus that required users to navigate multiple layers to access advanced functions.

By contrast, the interfaces in 2004 had become more streamlined and user-friendly. Microsoft Word 2003 introduced task panes, which consolidated various functions and provided context-sensitive help, reducing the need to navigate through menus (Microsoft, 2004). OpenOffice.org Writer focused on simplicity, offering a cleaner interface with customizable toolbars (Sun Microsystems, 2004). Corel WordPerfect 12 continued to refine its interface, emphasizing ease of use with improved icons and better organization of features (Corel Corporation, 2004).

3. Accessibility: Bridging the Digital Divide

In 1998, accessibility features in word processing software were rudimentary. Options for screen readers, magnification, and keyboard shortcuts were limited, making it challenging for users with disabilities to fully utilize these tools. Microsoft Word 97 did include some basic accessibility features, but they were not as comprehensive or user-friendly as modern standards (Microsoft, 1998).

By 2004, there was a noticeable improvement in accessibility. Microsoft Word 2003 included better support for screen readers and more extensive keyboard shortcuts, facilitating use by individuals with visual impairments (Microsoft, 2004). OpenOffice.org Writer, being open-source, benefited from contributions aimed at enhancing accessibility, such as improved compatibility with assistive technologies (Sun Microsystems, 2004). Corel WordPerfect 12 also made strides, incorporating more intuitive keyboard navigation and better screen reader support (Corel Corporation, 2004).

4. Performance: Speed and Stability

Performance is a critical factor in word processing software, impacting both individual productivity and overall user satisfaction. In 1998, the performance of word processors was constrained by the hardware limitations of the time. Microsoft Word 97 and Corel WordPerfect 8, while capable, could be sluggish when handling large documents or complex formatting (Microsoft, 1998; Corel Corporation, 1998). Lotus Word Pro, despite its innovative features, often struggled with stability, leading to frequent crashes (IBM, 1998).

By 2004, improvements in both software optimization and hardware capabilities led to enhanced performance. Microsoft Word 2003 was notably faster and more stable, thanks to better memory management and error handling (Microsoft, 2004). OpenOffice.org Writer, despite being a newer entrant, demonstrated impressive performance, particularly in its handling of large documents and diverse file formats (Sun Microsystems, 2004). Corel WordPerfect 12 also showed marked improvements in stability and speed, making it a reliable choice for professional use (Corel Corporation, 2004).

5. The Broader Impact of Word Processing Evolution

The advancements in word processing software between 1998 and 2004 reflect broader trends in the tech industry, including the push towards user-centric design, increased competition, and the rise of open-source software. These developments not only enhanced the functionality and usability of word processors but also democratized access to powerful tools for creating and managing documents.

In conclusion, the period between 1998 and 2004 was marked by significant advancements in word processing software. The evolution from the feature-limited, often cumbersome programs of 1998 to the more sophisticated, user-friendly applications of 2004 underscores the rapid pace of technological innovation. These changes were driven by a combination of user feedback, competitive pressures, and the overall growth of the tech industry, leading to software that was more powerful, accessible, and enjoyable to use.

References:

1. (1998). Microsoft Word 97. Retrieved from https://www.microsoft.com
2. Corel Corporation. (1998). Corel WordPerfect 8. Retrieved from https://www.corel.com
3. (1998). Lotus Word Pro. Retrieved from https://www.ibm.com
4. (2004). Microsoft Word 2003. Retrieved from https://www.microsoft.com
5. Sun Microsystems. (2004). org Writer. Retrieved from https://www.openoffice.org
6. Corel Corporation. (2004). Corel WordPerfect 12. Retrieved from https://www.corel.com

[1] That last one is levity … for now.

[2] Which one is which is debatable.

[3] Clippy was an animated paperclip that would pop-up in older versions of Microsoft word with helpful tips.

---

Dorothy Haraminac has performed traditional financial forensics and blockchain forensics for over a decade. She has testified as an expert on crypto tracing and value in multiple jurisdictions. She serves on several boards, has written extensively about crypto litigation, and is the country’s foremost expert on crypto tracing in civil litigation.  She was an adjunct professor at Houston Christian University where she taught Software Engineering, Cybersecurity and Digital Forensics courses. She holds a Master Analyst in Financial Forensics credential, holds a Certified Fraud Examiner credential, holds a Certified Cryptocurrency Investigator credential, has earned a master’s degree in Decision and Information Science, and is a licensed private investigator in Texas.

Ms. Haraminac can be contacted at (346) 400-6554 or by e-mail to admin@ybr.solutions.