Overview
This project was created as part of a Masters of Education course on Mobile & Open Learning by Jonathan Lapierre in 2020. As a teacher in British Columbia, Canada, Jonathan has seen the increased mobile use of his students as well as the increased desire for students to receive personalized and timely guidance and feedback at all hours of the day. This is particularly been the case during the shift to online learning during the 2020 pandemic. This site explores a future technology which looks to be a logical continuation of existing research and will transform how educators and learners access information and produce written work.
This resource has been created as a discussion-starter for educators to begin planning for the seemingly inevitable changes on their way. This future technology will be cross-platform, but will have the greatest impact when it becomes available to mobile devices. It is not currently a primarily mobile technology but in order to use its full potential it will transition. Its use will be ubiquitous, but in the context of educators the users will be students and teachers. As the work produced will be highly personalized to the user, users will require it to work no matter where they are, and the use of in-device camera and microphone is best facilitated in a phone or tablet format.
1 What is AI (artificial intelligence) writing?
AI Writing software uses the processing power of cloud-based servers with access to all digital written bodies of work to assist a user with many aspects of writing. This software when used on a mobile device will provide the best experience as the writing can happen when inspiration strikes, and conversations with the AI can incorporate local objects and locations. The software will also be personalized to the user's ability and incorporates phone hardware like the front and rear cameras and microphone. Some key criteria are listed below which describe the use and function of a future evolution of AI writing which will soon be in the hands of mobile device users.
User-based personalization - The user's writing style, age, and objectives are taken into account when producing written work.
Research help - The software can be given criteria and a topic and suggest arguments and provide sources for further research.
Ease-of-use - The software can be engaged with verbally on a mobile device for all aspects of writing and the software can interpret a variety of natural spoken language to adjust parameters on the fly.
Digital muse - users can input character traits and have live conversations with that character for inspiration.
Real-world inspiration – users can set up stories and characters based on nearby objects and locations through a discussion with the AI and AI characters with GPS as you walk.
Ideating - The user can explore essay and story ideas through conversation with the AI.
2 AI Data Sources
The AI writing software can examine all digital written work to not only pick up on subtle linguistic traits in order to better assist with work but can also evaluate the strength of your arguments. Access to journal articles, websites, books, and even podcasts will allow the AI to essentially access all available information on any topic.
This can be applied to a variety of vocations as well. Journalists reporting on developing stories in the field will also benefit from having access to information and live fact-checking by the software. The act of writing this new information adds to the AIs information bank, which updates its fact-checking in real-time. Consider the recent death of Representative John Lewis in 2020. As journalists report on his death, the software can suggest edits to anyone writing about John Lewis to include this new information. The software can even auto-generate biographical news articles for newsrooms.
“We are headed toward a situation where AI is vastly smarter than humans. I think that time frame is less than five years from now. But that doesn’t mean that everything goes to hell in five years. It just means that things get unstable or weird,” Elon Musk, July 2020.
3 Current Technology
AI is already being used in this way. Grammarly and similar sites use AI to suggest edits to vocabulary, clarity, and even tone.
AIWriter is a website which creates an article based on any headline you give it. I have demonstrated this by using the headline of Elephant Poaching and displayed the resulting AI-produced article here. The result isn’t perfect, but it already surpasses the work of many grade 10 high school students. This example opens the door to an interesting teaching opportunity. Similar in how students are taught to examine the sources behind Wikipedia articles, students can use the result from AIWriter to find sources and facts in the Cited Sources section. These facts and sources can help students select a focus for their own writing. AIWriter also has a “research-focused” mode which returns interesting information and concepts along with sources of further research back to the user based on their search query. There is a clear path of evolution ahead for this to become mobile-friendly and make its way into classrooms. Give it a try now at www.ai-writer.com
EssayBot is another AI writing website. When given a topic it provides choices, allowing for personalization of the message. This website directly takes quotes from sources and offers to rephrase them to avoid plagiarism detectors. The rephrasing often creates errors as it doesn’t choose synonyms that can fit in the context, but in just a few minutes a user can have the basic structure of a paper and associated citations. The way that it builds the essay section by section, allowing choice at each step gives the user more say in the end result. The user even has the option to write a few words that the AI will then use to build the paragraph. Give it a try now at essaybot.com
Google has also experimented with AI to see whether machine learning can be used as a tool by writers to inspire, unblock and enrich their work. As an educator, these these tools are stunning.
Banter Bot allows a user to chat with the character as they write the character. Imagine having elementary students who often struggle to create three-dimensional characters be able to interact with their character to find their opinions and reactions to events in the story. The character can be tweaked as they go on, and it even helps to build a backstory and motivations for the character.
Telescope brings the experience of Banter Bot into the strictly mobile realm. Imagine students taking their character for a walk in the real world and the app using GPS and map data along with objects recognized by the phone’s camera to combine the character’s attributes with the writer’s surroundings. This can generate story elements using the objects in the area. The “stream of consciousness” is vocalized as it is written into a Google Doc for later reference.
4 Future Development
AI Writing has progressed greatly in the 2010s, and the 2020s are looking to develop this ability beyond most human ability. Rather than replacing humans, the use of this technology as a classroom tool can prepare students for workplaces of the future and further push the transition towards inquiry-based learning. Students can use this tool to generate ideas and even to identify sources. Presently, these tools make errors in both readability and flow and they are not designed for mobile. A mobile AI Writing app will not only provide many of the features included above but allow users to receive targeted writing help at all hours of the day and provide a user experience akin to collaborating with a subject-expert. This can replace late-night emails from students the night before a task is due as the students can often work through writing blocks with the help of the AI. All tasks can be done verbally and the app can pose questions to the user verbally as well. The difficulty of language used can be adjusted for younger students or English language learners. Students can also work in their first language and have their work translated before submission. This mobile experience will even assist with creative writing, allowing users to take their characters on the go and build stories based on surroundings.
Using principles introduced in the UBC MET course of ETEC 523, the design of this mobile app will be crucial in determining how widespread it is adopted. The layout should use meaningful visuals, simple navigation elements with clear hierarchical choices in writing, and minimize extensive scrolling. The ability to quickly add images and format the text based on a number of templates can even see young students able to create picture books and tell stories.
5 Implications
The implication for students is that when a teacher assigns an essay on a topic the student can indicate this to the software verbally or by scanning the instructions with the device's camera. The student is then presented with a wide range of ideas, positions, and approaches to answer the topic. For research-based work, the student can then select more specific response information by examining the choices given by the software based on their interests and the facts backing up the argument. For creative-writing work, the app can work through creating a setting and characters and interact with the student to build the plot and add realism inspired by professional authors.
From a teacher's perspective, evaluation will also need to adjust away from the specific language use and paragraph flow and more towards the ideas, evidence, or plot selected by the student. Students will not always be selecting all the words but will instead be playing with idea selection and linking. Since all users will have access to the same information, selection of supporting details that the student believes to be most relevant will be of greater importance than the process of writing. This leveled playing field of information will also allow for more individualisation and inquiry-based learning as users will in effect have their own guide. This alone can have a huge impact on the future of education and the role of a teacher.
Also, this can spark new questions about the definition of plagiarism. Currently, the same AI tools can detect other AI-written documents, but this will likely get more difficult as they improve. Authors fluent in plagiarism issues have raised the question on authorship when AI is involved, and they equate it to a step further than paid ghostwriters. There are no clear-cut answers, but it’s clear that as more students access these features there is no simple way for a teacher to detect their use in all tasks and thus the methods and foci of evaluations will have to change. It may be simplest to force students to write under the supervision of teachers, but this would ignore a tool which can change the way we all access information. Inquiry-based learning is an effective strategy grounded in research which can be well-supported using technology such as outlined here.
The following clip discusses the future of AI and creativity.
Thank you for consulting this look into a future technology.
References
AI + Writing | Experiments with Google. (n.d.). Retrieved August 02, 2020, from https://experiments.withgoogle.com/collection/aiwriting
Artificial Intelligence and the Future of Plagiarism. (2020, August 01). Retrieved August 02, 2020, from https://www.turnitin.com/blog/artificial-intelligence-and-the-future-of-plagiarism
Bailey, J. (2019, February 12). How AI Will Change Authorship and Plagiarism. Retrieved August 02, 2020, from https://www.plagiarismtoday.com/2019/02/12/how-ai-will-change-plagiarism/
Bailey, J. (2019, November 14). Using AI Ethically While Writing. Retrieved August 02, 2020, from https://www.plagiarismtoday.com/2019/11/05/using-ai-ethically-when-writing/
Computer-Assisted Creative Thinking. (n.d.). Retrieved August 02, 2020, from https://www.csmonitor.com/1992/0916/16131.html
Generate unique text with THE AI article writer. (n.d.). Retrieved August 02, 2020, from http://ai-writer.com/
Marr, B. (2019, May 13). Artificial Intelligence Can Now Write Amazing Content -- What Does That Mean For Humans? Retrieved August 02, 2020, from https://www.forbes.com/sites/bernardmarr/2019/03/29/artificial-intelligence-can-now-write-amazing-content-what-does-that-mean-for-humans/
Pedaste, M., Mäeots, M., Siiman, L. A., Jong, T. D., Riesen, S. A., Kamp, E. T., . . . Tsourlidaki, E. (2015). Phases of inquiry-based learning: Definitions and the inquiry cycle. Educational Research Review, 14, 47-61. doi:10.1016/j.edurev.2015.02.003
Using an AI to Spot AI Writing. (2019, November 26). Retrieved August 02, 2020, from https://www.plagiarismtoday.com/2019/11/21/using-ai-to-spot-ai-writing/
Vincent, J. (2019, February 14). OpenAI's new multitalented AI writes, translates, and slanders. Retrieved August 02, 2020, from https://www.theverge.com/2019/2/14/18224704/ai-machine-learning-language-models-read-write-openai-gpt2
Write your best with Grammarly. (n.d.). Retrieved August 02, 2020, from https://www.grammarly.com/