Ten years ago, it was common for employers to ask candidates about their time management skills or proficiency with Excel. Those questions helped hiring managers determine whether a candidate could handle the role’s day-to-day tasks. But that’s changing now. As AI continues to integrate into the workplace, it’s becoming more and more important to evaluate candidates’ ability to work with it.
Traditional interview questions won’t be enough going forward
Traditional interview questions will still serve a purpose, but they won’t reveal whether a candidate knows how or when to use AI tools. AI-focused questions aren’t meant to assess technical AI skills, such as software development or machine learning; rather, the goal is to evaluate their non-technical AI skills, such as ethical discernment and output verification. Asking questions about these skills gives you a glimpse of how candidates already use AI and their understanding of best practices.
The best interview questions to ask about AI
I’m not saying that you’ll have to completely re-engineer your interviewing approach or focus only on these questions. Asking about your candidate’s strengths and weaknesses or where they see themselves in a few years still provides useful insight. You just need to make sure you’re also asking questions that reflect the modern world of work.
1. How do you decide when to use AI for a task?
This question highlights a candidate’s judgment when working with AI. They may describe using it to summarize information or brainstorm. Strong candidates should be able to explain how they evaluate tasks before turning to AI. Candidates should also be able to recognize when using AI isn’t appropriate, like in situations that involve sensitive data.
Make sure to clarify that the goal of this question is to understand how candidates work with AI tools, not whether they avoid them. Without clarification, some candidates may downplay their experience if they feel like they’re being tested on whether they over-rely on AI.
2. Describe a time when AI gave you a result that was incorrect or misleading.
AI tools can produce responses that are incomplete, irrelevant, or outdated. This prompt helps you determine if a candidate understands this risk and how to handle it in practice.
Candidates should be able to describe how they recognize when an output is wrong and what steps to take. Their method may include checking reliable sources or adjusting their prompt. Look for candidates who treat AI-generated work as a starting point rather than the final product.
3. If we introduced a new AI tool, how would you approach working with it?
Since AI tools evolve quickly, adaptability is just as essential as experience. Answers to this question provide glimpses into how candidates approach unfamiliar technology.
Ideal candidates will start with low-risk tasks to determine how the tool performs. They may mention experimenting with different prompts and identifying situations when the tool improves efficiency. Candidates should also focus on learning the tool’s strengths and limitations. The goal isn’t immediate perfection. The goal is to see if the candidate provides a thoughtful approach to testing the new tool.
4. How do you protect confidential data when using AI tools?
Protecting confidential information is a critical part of responsible AI use. While many employees understand general data security rules, fewer consider how those rules apply to AI systems.
Candidates should acknowledge that sensitive information shouldn’t be entered into AI tools. They may mention removing identifying details or proprietary information. Another good response is following the company guidelines that outline what can and cannot be shared with AI. Not every candidate will have experience working with a formal AI policy, but they should be able to demonstrate that they understand certain information must be protected.
Should the candidate have certifications or formal training?
For most roles, certifications or formal training aren’t always necessary. I suggest opening up the questions to how candidates use AI on their own time or in other projects.
The Pew Research Center found that 31% of Americans already use AI several times a day. So some candidates may already perform these skills without realizing there’s a name for it. For example, they may know to avoid entering their personal information into AI tools or remain skeptical of AI outputs until they investigate further.
Is it okay to hire candidates who don’t have these skills at all?
Yes, but consider where your company is in the AI adoption journey. If AI is fully integrated into your workflows, you may need new hires who understand how to work with it from day one.
If you’re still in the implementation and experimentation phase, however, it’s okay to hire candidates who may not have a ton of experience with AI yet. The new hires can develop the necessary skills alongside the rest of your team. In this case, factors like a positive attitude, a willingness to learn, and overall cultural fit matter more than experience.
