I don’t have Instagram. No Snapchat, no TikTok, and my phone has maybe six apps I actually use. I was late to smartphones, later to social media, and I’ve mostly stayed that way on purpose. If a tool doesn’t clearly make my work or my life better, I don’t need it cluttering either one.

So, when people ask how I feel about AI, the honest answer is: about the same way I feel about every other tech trend. Skeptical by default, willing to be convinced,
in no hurry to be first.

But AI is different from TikTok in one important way, I don’t get to opt out of it. Clients need it and the competition and candidates are using it. The tools I rely on every day are quietly building it in whether I ask for it or not. The question isn’t “should I use AI,” it’s how to use it deliberately, with a clear sense of what it’s for and still feel ok about it.

Even the people building it are nervous

I’ve spent some time looking into what people inside the major AI labs are saying, researchers and executives who’ve actually built these systems. It’s sobering. Just this week, Jacob Coxon, who worked at both OpenAI and Anthropic, resigned publicly, saying both companies were “racing straight to self-improving superintelligence and gambling with our lives.” A steady stream of safety researchers and other employees have left major labs over the same worry. Anthropic’s own alignment lead, Evan Hubinger, publicly backed Coxon’s statements, saying he believes there’s a real chance AI could go catastrophically wrong within a decade, more specifically, “We really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade.”

That’s a very different level of concern than whether I should let AI draft a job posting — a distinction that matters more than it usually gets credit for when talking to AI cynics. Those warnings are about the trajectory of the technology itself — companies racing to build systems more capable than they can control. It’s not the same conversation as whether I, at my desk, should use AI to clean up a search string or draft an outreach email. Treating every chatbot use like a step toward some larger danger doesn’t hold up, and we skeptics need to remind ourselves of that.

A third layer worth watching

There’s another risk that gets talked about far less: what happens to the information we type in. PagerDuty Shadow AI’s recent survey found 88% of office professionals have shared work information with public AI tools, and nearly a third shared confidential documents or financial data. It’s a scenario anyone in recruiting will recognize: uploading a resume or a client’s salary band into a consumer AI tool to speed up a comparison. Company leaders often don’t see how far this has already spread — McKinsey’s “Superagency in the Workplace” research found C-suite leaders estimated only 4% of employees use generative AI for a significant share of their daily work, while employees reported the real figure was closer to 13%. In our work space, it certainly feels like twice that number, at least.

For a recruiter, that’s not abstract — we handle exactly the kind of information this problem is made of: resumes, comp details, hiring criteria shared in confidence. Pasting any of it into a consumer tool without thinking is a small decision a lot of us are probably making without realizing it’s a decision.

So — how do we actually do this?

I don’t think the answer is refusing to use AI, any more than my answer to social media is refusing to own a phone. It’s the same instinct I apply to everything else: use it where it clearly helps, understand what it’s doing before you trust it, and don’t hand it something you wouldn’t want floating around outside your control. In practice:

• AI handles mechanics, not identities. Search strings and first drafts, yes. But not a candidate’s resume or a client’s comp band without knowing what tools are company approved and IF any of your company tools are approved for sensitive, PHI and regulated data.
• Understand Regulations. AI in hiring is one of the most regulated AI use cases (NYC Local Law 144, EEOC guidance, Colorado’s AI Act). For myself and our clients, this becomes a critical governance issue when developing and using sourcing and screening tools, agentic prioritization etc. For our consultants, every new AI project will require staying up to date on client specific compliance and regulations in this space.
• Know the difference between a free tool and a business one. It changes what happens to what you type in; will it train your model? How long it keeps the data? Is this an Enterprise agreement? These are the questions Logisolve can help your team understand and answer.
• Stay honest about the bigger picture without letting it paralyze the smaller decisions.

I’m not going to be first in line for the next AI trend any more than I was for Snapchat. But caution and competence don’t have to be in tension — use what helps, understand it before you trust it, and don’t let hype make the decision for you.

This is also, increasingly, my job — you knew there was a pitch coming

A growing share of my sourcing work these days is for AI-related roles across the board: engineering, product, and leadership positions where a company is building an AI team. This isn’t purely abstract for me, I am learning and evolving with this space every day.

If you’re working in AI, know someone who is, or have even heard of someone who may be good for AI-focused engineering, product, or leadership roles, we always welcome a conversation, referral or even a name — including a referral fee for anyone who sends a placed candidate my way.

(Full disclosure: this article was written with the support of AI 😊)