AI-assisted writing: Friend or foe?

Communication, especially written communication, is one of the most important abilities in almost any career. Being able to effectively convey your ideas and instructions can improve your working relationships and avoid misunderstandings. Fortunately, there is a plethora of resources that can help improve written text—and even generate it for you. As the tools get more sophisticated and can perform more-complex tasks, however, it becomes necessary to critically evaluate how they are, and should be, used.

Spelling and grammar checkers.

Spelling and grammar checkers have been around a long time and have significantly improved many people’s writing. But there are entire websites devoted to the often humorous outcomes of incorrect fixes. Most people have learned to double-check the automated tool’s suggestions and add commonly miscorrected terms to their personal dictionary. These tools are designed to catch obvious errors.

Tone and authoring.

Newer tools offer suggestions to shorten text, change the tone of an entire message from formal to friendly, or make the author appear more confident. The tools go beyond proofreading and copyediting to ensure that text follows established rules of good writing and into the territory of substantive or developmental editing. As a result, the document flows logically, transitions between ideas are clear, and there is no extraneous or distracting information. These types of changes are significantly more substantial and subjective, often involving reorganization or rewriting of entire sentences or paragraphs.

Artificial intelligence tools (often called generative AI) can create “original” text that reflects their training set or user input, and they can be used to help with a job search. The tools can create an email template that requests an informational interview or write a letter of recommendation. Using a LinkedIn profile or job description, the tools can also write a résumé or cover letter. But how much should they be used?

Ethics.

As these tools become increasingly complex, the ethical questions surrounding their use become more difficult. The suggestions in “Best Practices for Using AI When Writing Scientific Manuscripts,” published recently in ACS Nano, can also be applied to AI’s use for career documents.

▸ Use AI-generated text to break writer’s block, spark new ideas, and push your own thinking in new directions.

▸ Treat anything generated by AI as a rough first draft; check and double-check everything for accuracy—including references. Some tools have been known to create false references to support inaccurate conclusions.

▸ Remember that the tool is not an author and cannot generate new ideas, only regurgitate what it has read elsewhere.

▸ Do not use the generated text verbatim, as it may have been copied word for word from another source.

▸ Make clear to the reader how much AI was used in the creation of the document, as well as how it was used.

As technology improves, people will learn to adapt and come to a consensus on appropriate use. But in the end, a human author is responsible for everything that is created and distributed as their work—no matter how much technological help was involved.

Get involved in the discussion. The ACS Career Tips column is published monthly in C&EN. Send your comments and ideas for topics for future columns to careernavigator@acs.org.