This post was originally published on this site.
Allison Harbin jokes that she started working in artificial intelligence by accident.
Harbin earned a doctoral degree in art history from Rutgers University in 2017 and spent much of her twenties and thirties working in higher education, intending to become a professor.
Then, in 2023, she was recruited to work on Gemini, Google’s generative AI chatbot, designing and refining its search engine responses. She joined the project in October 2023, figuring it would be an exciting opportunity to try something new.
Harbin’s contract lasted six months but made a lasting impression on her career: Harbin has worked in AI ever since, a field that she describes as “equally challenging, exciting and fulfilling.”
She joined the wave of a new kind of AI engineer — one without formal technical training or a bachelor’s degree in computer science. Known as AI prompt engineers, anyone from marketing executives to lawyers can qualify for the buzzy job.
Prompt engineers are the key players behind how AI models like ChatGPT, Gemini and Microsoft Copilot generate accurate, relevant responses.
By crafting precise inputs, they ensure AI delivers the best results in real-time interactions, streamlining communication between users and technology. As AI becomes increasingly embedded in various industries, the demand for prompt engineers is skyrocketing, with salaries reaching up to $300,000.
Some economists have cautioned that the sudden demand for prompt engineers could be a flash in the pan, while other technologists have predicted that prompt engineering could soon evolve from a job title to a skill that is expected — or required — of desk workers.
Those who are working on designing and refining the technology, like Harbin, are betting on AI’s staying power.
For the past nine months, Harbin, now 38, has been working as an AI analyst and prompt engineer at a small health-care tech company. She works remotely from her apartment in Chicago and earns over $100,000 per year (she declined to share the name of her employer or exact salary).
Harbin coaxes AI models like OpenAI’s GPT daily, utilizing natural language to encourage the model to produce clearer, smarter responses. She also teaches users how to use the generative AI tools her employer designs for their daily work tasks.
Here’s how Harbin got into prompt engineering and what a career in AI really looks like:
CNBC Make It: How did you decide to start working in AI?
Allison Harbin: If you had told me five years ago that I would be working in AI, I would’ve been like, “What are you talking about?” However, I was stitching together a living writing white papers and assisting on different research projects when ChatGPT came out in 2022. And, whether it was a coincidence or not, I lost a big client shortly after its release, and a recruiter reached out with the opportunity to work on some of the writing and communication elements of Gemini, Google’s AI chatbot.
I don’t have a coding background and I’m also not a data scientist, but when I started working in prompt engineering, I discovered that a lot of the skills you need for the job are just clear communication and effective writing, which I can do well. The more I did it, the more I enjoyed it — it’s like there’s a hidden code waiting to be cracked.
Which resources did you use, to close any technical skills gaps as an AI prompt engineer, since you don’t have that coding background?
I subscribed to LinkedIn Learning [which costs about $30 per month] and completed a couple of relevant certificates there, like “Ethics in the Age of Generative AI” and “Advance Your Skills as a User Experience Researcher.” But there are also a lot of free resources you can take advantage of. Don’t underestimate what you are capable of figuring out on your own, you just need the time and diligence to snoop around on the internet.
What does your day-to-day look like now as an AI analyst?
This is where I think not having a formal tech background comes in handy. A big part of my job is training and showing people how to use the robot, or GenAI chatbot, that my team has designed. It’s intended to help employees in the professional services sector do their jobs more efficiently.
What is your favorite part of the job, and what is the most challenging aspect?
My favorite part of my job is talking to people about how they want to use AI or generative AI, and then figuring out if it’s possible. It’s like solving a complex puzzle — if it is possible, I focus on crafting the right prompts to get the best outputs. The most challenging part of my job is managing the rapid pace of change in the field. It’s like drinking from a fire hose, and that can be overwhelming at times.
What’s one of the biggest misconceptions you’ve heard about working in AI?
A common misconception I’ve heard is that AI will completely replace jobs. While some tasks may change, I think humans will always need to be in the loop to proofread, fact-check and train AI. The most common question I get is, “Am I going to lose my job to AI?” My answer is probably not, but certain aspects of your role might evolve. The technology isn’t perfect — sometimes the data it provides can be inaccurate, and it may even produce hallucinations.
When I run training sessions, I talk a lot about how many large language models are trained on data from the open internet, where not all sources are reliable. Some can be biased or inaccurate. It’s important to be aware of these biases.
What is your biggest hope for AI, and your biggest fear?
I think my biggest hope for AI is that as we learn how to use this technology for work, we also create new jobs around it, and we’re not just using it to automate people’s jobs. As for fears, I think this rush to adopt and create new AI technologies without properly understanding what these tools are capable of and where they fall short is something we all need to be aware of. Ultimately, I hope we catch any mistakes or problems AI can create before they become too big to fix.
This interview has been edited and condensed for clarity.
Want to up your AI skills and be more productive? Take CNBC’s new online course How to Use AI to Be More Successful at Work. Expert instructors will teach you how to get started, practical uses, tips for effective prompt-writing, and mistakes to avoid. Sign up now and use coupon code EARLYBIRD for an introductory discount of 30% off $67 (+ taxes and fees) through February 11, 2025.