Generative Artificial Intelligence can now be accessed by virtually anyone, not just through dialogue-type prompts in well-known models, such as ChatGPT, but also through AI-enabled tools like Notion AI and Readwise Reader, which offer users user-friendly interfaces to take advantage of multiple levels of automation.
Some of the most useful ways in which AI can help linguists enhance their productivity are terminology extraction and glossary creation, parallel text creation, document summarization, context research and examples, extended definitions, and more.
Translation is of course a function AI is capable of handling in hundreds of languages, but this article focuses on all tasks linguists perform as part of a translation project and which can be streamlined with AI.
Although not all AI models can perform all the tasks listed above, at their core is generative AI, which is a type of artificial intelligence that creates something new based on what it has learned (i.e., what it has been “trained on”) just like writers draw inspiration from observation, knowledge or experience to write a story.
Some of the tools that can do most or all the productivity-enhancing tasks listed in this article are the subscriber (paid) versions of AI-enabled systems like ChatGPT, Microsoft Bing (Copilot), Notion AI, and Readwise Reader.
A few free tools, like Google’s Gemini, can also help linguists get a lot done, fast, but it is important to note that only the paid versions of most solid AI tools preserve confidentiality or give users the option to prevent their data from being used for model training.
Writing. AI can suggest ideas or relevant research articles/papers to give users a start. AI can also provide different perspectives when prompted to adopt a persona. For example, linguists can ask it to write a cover letter based on a job post “in the style of a news anchor,” and see what develops.
AI can also help writers move forward when stuck. It is possible to give AI a sentence and prompt it to continue writing. Then, users can validate and edit the output. This can save a lot of time.
Changing style and tone are other capabilities of AI. For example, if a writer is unhappy with the way something reads or if the intended target audience has changed, the AI can be prompted to change style and tone based on the target audience. For example, the AI can be told to “rewrite the following text in an informal, cheerful style appropriate for teenagers.”
AI can also provide definitions, find synonyms, check grammar, clarity, conciseness, and analyze other factors to suggest improvements.
Terminology Extraction and Management. AI allows linguists to pull terms from a document, image, or other source. In most AI apps that require subscriptions, users are able to attach entire files and prompt the AI to extract domain-specific terms. For example, “Find all the terms related to [subject/domain] in the following [file] and create a list showing how many times they occur: [attach file].”
Linguists can create a list of terms in any language using simple prompts like “Create a list of XX terms about [topic] in [language], and [add details to get more targeted results or ask the AI to add definitions and format the results in a specific way, like a bulleted list or a table].”
To add context to a list of terms, the AI can be given a prompt like “For each of the following terms, add [XX] example[s] of how they are used in a [domain/geography] context. Give examples in [language].”
Creating a glossary once terms are extracted can be done fairly quickly by asking the AI to translate terms. Write a prompt to ask the AI to translate a list of terms into your choice of languages. For example, “Translate the following list of terms into [language].” You can also prompt the AI to organize the terms in a table and sort the terms in any particular way, not just alphabetically.
Certain AI apps, including Notion, allow users to prompt the AI to align texts at the sentence or phrase level text. If aligning texts, to get the best results, prompt the AI to “create a bilingual table for [text] in [languages].”
If doing research for a translation project, readers can save content for later using functions like those in Readwise Reader’s Ghostreader. This app allows users to import web pages, documents, and media and store them for later use.
Summarization is another very popular way to get through research tasks faster. Just about all subscriber versions of popular AI tools have the capability to summarize entire documents based on a prompt. The more detailed the prompt, the more targeted the summary. Linguists can also summarize just sections of documents. The key is to create thoughtful, purpose-clear prompts so that AI generates as close to a result as expected.
Different versions of the same text: Gemini (formerly Bard), for example, will create three different versions of a summary or answer requested. It will also give users the option to validate the answer chosen based on an Internet search and provide sources. User beware, though, AI can lie (hallucinate), and all information must still be human-fact-checked.
Finally, general text analysis is another time-saving capability of most AI tools. For example, in Readwise Reader, linguists can upload an entire document and ask the AI pointed questions about it, such as “Does this document address Regulation XYZ?”
Linguists are advised to test a few apps and see how much those apps can simplify and automate their day-to-day tasks. At this time, not one single tool can provide everything an individual is looking for, and experimentation is actually recommended to see what works best in every case.
