literature review ai

literature review ai

The Impact of Artificial Intelligence on Literature

1. Introduction

Artificial intelligence is a broad topic, and its effects are hard to understand within a short essay. The effect that AI will have on literature will differ in its form, and how greatly it will impact the way literature is produced and critiqued. Thus, the aim of this essay will have to narrow the scope of impact on a specific type of literature, that being interactive digital literature. The hardware and programming behind AI is very complex and is growing ever more powerful. Hopefully the focus of the essay will not get disparate of the overall effects, by focusing on just interactive fiction can still provide a broad analysis. As to the overall effect that AI will have on literature, this is also a difficult question and its ramifications will likely not be fully understood for some time. Some effects will be subtle, others profound. Interactive fiction itself is a (relatively) small subgenre of literature and so expecting great change will be presumptuous. However, interactive fiction has often been at the forefront of experimental methods of approaching literature and has sometimes even had an impact on more traditional forms of storytelling. Often changes in the creation and distribution of literature technologies have been hard felt within the walls of even the most traditional literary institutions. So we must also consider the indirect effects of AI on literature.

2. Current State of Artificial Intelligence in Literature

A very large portion of the work on fiction from AI applications was produced in the 1980s and early 1990s, mainly in the form of short stories or prose poetry that imitates some elements of surrealism, expressionism, or symbolist poetry, given the ease of simulating the unstructured form of these genres. These programs often used their ability to generate large volumes of text-based output as a measure of their success as creative systems, but for the most part, the writing quality was poor. The later assistance work of the Brut writing simulator, and a very large number of algorithmic and more traditional approaches to automated story generation also fall under this banner. Currently, there are several AI projects that produce outputs that could be called narrative fiction of a broadly literary kind. There are projects to develop automated joke and anecdote telling systems, the most notable of which is MARGIE, the automated stand-up comedienne, but these are not usually aimed at generating written output. Of specific interest today are the story generation efforts being undertaken in several academic AI groups, using approaches that range from tightly structured deep linguistic analyses of narrative to more shallow approaches based on predefined story representations. Typically, the stories generated are very short and simplistic, due to the difficulty and computational cost of large-scale narrative generation.

3. Benefits and Advantages of AI in Literature

First let’s pretend that robots can think and communicate as well as we do. First of all, programming rote tasks like grammar checking and citation checking into AI is a very viable option. Sure, anything an English major can do can be done on a word processor like MS Word. Other than the fact that Word’s interface is cluttered, unintuitive, and prone to random crashing, it still lacks many useful features. Something like grammar checking in Word is unreliable at best. Sure, in the most basic sense it can catch run-on sentences and subject/verb agreement. Still, put a sentence with a complex structure that is still correct grammatically, but if a couple of words were left out, would be very difficult to understand. As it stands now, Word would suggest breaking the sentence into pieces and capitalizing the first letter of the words following the ones it thinks should be a new sentence. A grammar checking AI could analyze the sentence and the context and suggest adding in the missing words. To attempt the more ambitious goal of making AI that can write literature, we need to go much further. Though an AI could be given a large database of text to use as sample material and be set to mimic the styles used, this still takes away from the creativity that is involved in literature. At its height, creative writing is an original synthesis of information to present new ideas, stories, and opinions. While I’m not discrediting the authors of SpongeBob, for example, there has been an AI (the episode was actually about AI) that was programmed to imitate a script. With AI, we could actually take text mining to the next level. Data mining is a technique used to retrieve information from a database. This can vary from literally any database with information and is often used for marketing purposes. Text mining is a specific use of data mining, primarily used for matching database information to a collection of text or its description. This would be extremely useful in all research domains. For example, an AI could be let loose on the internet to find information about an obscure rare disease for which it has a symptom list, and the doctor needs to find a diagnosis. Coming back to our own field, text mining has been used to analyze literature for useful patterns. In the past, this has been mainly for the purpose of making critiques on style or content, grading, and compiling dictionaries. An AI’s ability to actually discover new information in text and understand would take text mining to another level. Imagine an AI reading a collection of historical documents and piecing together a more accurate account of an era or tracking changes and influences on the English language by comparing works of literature from different time periods.

4. Ethical Considerations and Challenges

The development of any new technology for the purposes of arts and culture always necessitates difficult ethical and moral questions. The Greek myth of Pygmalion who creates a statue, which Galatea brings to life as a young woman has been repeated by authors and artists for centuries. The idea of creating a living, breathing form out of an inert material is a dream shared by many. This process is nearer to a reality than one might imagine due to advances in AI and techniques to simulate human thought and appearance. As artificial life becomes more believable as a living entity, the question of AI rights and the distinction between the real and simulated will be a point of contention. An artificially intelligent simulated person may be created to serve as an actor in a game or a story, only to be subjected to the desires of its creators and players. They may be used as a tool or a slave, never possessing any true rights or free will of their own. This is a grave moral issue, for even the AI which does not possess true life may serve to reinforce attitudes and actions which are harmful to living people of certain creeds. For the question of what constitutes a living person is not one that is easily answered and many may come to feel genuine empathy for an artificial entity in the future.

5. Future Prospects and Implications

At its very worst, the machines will simply remove the human writer from the literary process. Human writers will be unable to compete with the speed and cost of AI production and will not be able to find employment in a profession overrun by machines. This could mark the start of a new dark age for literature, with machines churning out an endless deluge of formulaic pot-boilers with no artistic merit. AI literature may reflect the postmodern condition of humanity’s ever-diminishing quest for meaning and truth.

Many have made the argument that as the AI becomes more sophisticated, it will develop the ability for independent thought and action. At this point of machine autonomy, society will be faced with a number of ethical questions on the use of AI. Should machines reach this point, then it is possible that they will become writers in their own right; creating works of fiction about the experiences of a nonexistent human society.

Al will only become more prevalent and embedded in society. In terms of literature, this could mean an overhaul of the entire profession. Some writers will no doubt welcome this change, arguing that the AI process will enable them to concentrate more on the creative aspects of writing. However, the comments made by professional songwriters on the AI music program suggest that the entire writing process may become so clinical that the actual writing becomes a mere afterthought.