sompong_tom/istock/getty images plus
All of a sudden, we appear spoiled for choice in terms of being able to access AI language models that seem smart enough to write essays, plan out exercise and diet plans, create weeklong agendas for professional development conferences, or even write your next résumé and cover letter. But is it helpful to leverage such tools as part of a job search?
Well, yes, absolutely—and also no. Actually, it will depend on how you are using these platforms and how you can integrate their immense brainstorming abilities as part of your own efforts. Having spent time playing around with some of the AI tools, here are some of my recommendations for how they can support graduate students exploring career options and applying for opportunities that they’ve identified.
Let’s first address some important assumptions about these AI models. They are still in their infancy, and the content that they provide—from wherever they acquire it, with or without permission—should not be assumed to be correct. In fact, I would go on the assumption that every statement involving facts and data is a lie—a confident, well-spun lie.
Let me illustrate. I asked ChatGPT, “What are some specific data (with citations) about how first-generation graduate students can benefit from networking?”
Here is part of the response:
“First-generation graduate students can benefit significantly from networking to establish professional connections, find mentors and access career opportunities. Here are some specific data and citations to support this:
- First-generation graduate students face challenges in networking due to limited social and cultural capital, which can affect their career advancement opportunities (Kim et al., 2019).
- Kim, Y. K., Lee, K. H., & Song, J. H. (2019). First-generation graduate students’ social and cultural capital and their networking behaviors. Journal of Diversity in Higher Education, 12(3), 220-234.”
This looks great, except that this is a completely made-up citation in a real-life journal. It doesn’t exist. It sounds like it could, or should, but it doesn’t. When I responded by saying this, I got this response:
“I apologize for my mistake. As an AI language model, I do not intentionally make up citations, but I can sometimes generate incorrect information. Here are some actual references to specific data about how first-generation graduate students need to network:
- First-generation graduate students face unique challenges in networking due to limited social and cultural capital, which can impact their career advancement opportunities (Bhattacharya & Williams, 2019).
- Bhattacharya, G., & Williams, J. (2019). From the margins to the center: Exploring networking experiences of first-generation graduate students. Journal of Student Affairs Research and Practice, 56(4), 457-470.”
And this is also a completely made-up citation in a real journal.
With this understanding of the limitations of AI tools, their benefit comes not from generating original content, like a whole new résumé or cover letter, or possible answers for common interview questions that you might get. Rather, they come from helping you reflect on your pre-existing materials or brainstorming with you some next steps take in your career-development process.
So, here are some questions you could pose as part of an exploration of different career paths:
- What are some of the keys steps I should take when exploring different career paths as a graduate student/postdoc?
- Provide a detailed two-year timeline for exploring careers and gaining relevant experience for a Ph.D. student in the humanities. (And then follow up with more prompts: “Provide even more detail and create a way to track progress.”)
- What are some of the skills valued in ABC industry?
- What does the term “XYZ” mean when used in descriptions for biotech research jobs (e.g., cross-functional, matrixed environments)?
Résumés and Cover Letters
When you are in the job application phase, you can copy the text from a job description and ask the AI tools to identify the main skills or knowledge areas sought. You would want to combine the answer you get with real-world information you have gained from your networking just to be sure, but these tools are great at identifying trends and patterns in data. You can then cut and paste the content of your résumé (removing your contact information and anything sensitive you don’t want to be shared, copied and regurgitated) into the AI tools and ask what the main themes are from your résumé. Ideally, the main themes from your customized résumé should overlap with the main themes from the job description. This is a more nuanced version of matching keywords from the job description in your résumé, because it is focusing more on the context of the skills and knowledge that make up the keywords (similar to a what a hiring manager would do and more advanced that what most applicant-tracking software does).
Feeling stuck with writing a cover letter? You could start off by asking the AI tools, “How should I structure a cover letter for ABC type job in XYZ industry?” You are not just going to blindly follow the answer you get, but some of the ideas can help you get started and prompt your own thinking as well.
Additionally, you can paste the text from a first-draft cover letter into an AI tool and then ask it to make that draft sound more professional. The content is yours (and this takes a good amount of time to get into a first-draft state), but you can borrow ideas from the AI tools about how to present it best in your cover letter.
This approach is especially helpful for international students who want their cover letters to read more naturally, allowing the reader to focus more on the content (and all the great experiences, skills and knowledge that they have to offer). I wouldn’t use the suggested text exactly—sometimes it won’t be an improvement—but it can often help provide some good ways of sharing the information.
This raises the question of the ethics of this approach. In my role providing career advice to graduate students and postdocs at the University of Pennsylvania, I give people suggestions on how they can tweak the content and formatting of their application materials. I don’t write the résumés or cover letters—my feedback is for people to think about and implement in whatever way feels helpful. The original content belongs to the student or postdoc, and the final version might include a few of my recommendations. Getting feedback from an AI on a cover letter doesn’t really differ from that in too many ways.
If you are looking to identify the type of questions that you might face in a job interview, your networking will be a key part of gathering this information. However, you can certainly ask the AI tools lots of questions, like the following, about interviewing as a small part of your preparation.
- What types of questions are commonly asked in screening interviews for ABC roles?
- What questions can I ask hiring managers as the end of an interview for XYZ role?
- How can I leverage my experience doing archival research in ancient history to answer a job interviewing question about being a good problem-solver?
The answers you get are not going to be something you can use just as they are in your next interview, but they will give you an opportunity to reflect and brainstorm about what questions you might expect, as well as what stories you can tell from your own specific experiences to answer them.
In conclusion, to finish this post, I couldn’t resist asking the AI model to create a summary paragraph based on what I have written above. Here goes:
“In conclusion, while AI language models have made incredible strides in generating content that can be helpful in many aspects of career development, it’s important to remember their limitations. Rather than relying solely on AI-generated content, the real value of these tools comes from using them to brainstorm and reflect on your pre-existing materials. From exploring different career paths to refining your résumé and cover letter, these tools can help jump-start your creativity and give you a fresh perspective. So don’t be afraid to try them out and see what kind of new ideas and insights they can bring to your job search!”
That’s not too bad, but mine would have been better.