In a recent LinkedIn post, Callum Laing discusses the significant, yet often overlooked, risks associated with the rapid adoption of Artificial Intelligence (AI) tools, particularly concerning confidential information and data breaches.
Laing highlights the paradox of businesses encouraging employees to use AI for productivity gains while simultaneously exposing sensitive data without adequate safeguards. He points out that this practice, while potentially boosting output, creates a substantial vulnerability, especially when customer data is involved.
“AI is a fantastic tool to help people become more productive. But we’re all enabling our junior staff to put confidential information into tools with no oversight. In fact, we’re demanding it!”
The Growing Threat Landscape
Callum Laing emphasizes that the problem is not theoretical, citing the financial sector as a prime example. He references a Forbes report indicating a significant increase in AI usage among Family Offices, which manage substantial global wealth. This trend, while driven by a desire for efficiency, introduces new vectors for potential data compromise.
Laing anticipates that while major corporations with vast amounts of customer data might be the focus of initial, high-profile breaches, smaller financial institutions will also face severe consequences. He predicts these smaller entities will struggle to contain the damage after the fact, with their difficulties largely remaining out of the public eye.
“The obvious solution is to run your own inhouse LLM, but we’ve spent the past decade outsourcing our hosting / networking knowledge to the cloud. And with new, better, models being release every few weeks nobody wants to get left behind.”
The Dilemma of In-House vs. Cloud AI
A key challenge identified by Laing is the practical difficulty of implementing secure, in-house AI solutions. He notes that a decade of reliance on cloud services has diminished internal expertise in hosting and networking. This makes the prospect of managing a private Large Language Model (LLM) daunting for many organizations.
Furthermore, the rapid pace of AI development, with new models emerging frequently, creates pressure for businesses to stay current. This continuous innovation cycle makes it challenging for companies to commit to and maintain a specific in-house solution, as they risk falling behind quickly.
The Unseen Impact on Smaller Firms
Callum Laing argues that the true extent of AI-related data breaches may be underestimated because smaller firms, often operating with less public scrutiny, will be disproportionately affected. These breaches, while potentially devastating for the institutions involved, may not make headlines.
“Interestingly, the data breeches you are going to hear about, will be the big companies with the most retail data, but I can absolutely guarantee you, there are going to be a lot of small financial institutions that will be desperately be trying to slam the AI door shut, long after the horse has bolted.”
Laing concludes by framing the situation as a complex dilemma, or “quite the pickle,” for businesses navigating the dual demands of enhanced productivity through AI and the imperative of robust data security. His analysis underscores the need for a more cautious and strategic approach to AI integration, particularly concerning sensitive information.
📝 About This Content
This article is based on insights shared by Callum Laing on LinkedIn.
📅 Originally posted on June 22, 2026 | View original post on LinkedIn →