Generative AI in the Workplace: Simplicity or Illusion?

Imagine a workplace where routine tasks that once required hours are accomplished in mere seconds. Morgan Stanley wealth advisors now rely on an AI tool called Debrief to turn lengthy client meetings into structured summaries, follow-up emails, and CRM updates almost instantaneously. Similarly, Klarna, the global fintech company, uses generative AI to handle millions of customer chats monthly, significantly improving efficiency and customer experience. At first glance, such technologies seem to offer businesses the perfect trifecta: broad capability, high accuracy, and remarkable simplicity. But is this simplicity genuine or just an illusion?

Understanding the GAS Framework

The real impact of generative AI in the workplace requires understanding the Generality-Accuracy-Simplicity (GAS) trade-off, a concept explored by researchers Sharique Hasan, Alexander Oettl, and Sampsa Samila in their fascinating June 2025 paper "From Model Design to Organizational Design: Complexity Redistribution and Trade-Offs in Generative AI", available here. This paper applies the GAS framework (from social science research) as a contemporary adaptation of Thorngate's postulate of commensurate complexity, a foundational principle stating that no theory can simultaneously be general, accurate, and simple. According to this postulate, enhancing one or two dimensions inevitably requires compromises on the third. In our context:

• Generality - the breadth of contexts or tasks the AI can handle.

• Accuracy - how reliably the AI produces correct or appropriate outputs.

• Simplicity - the ease with which users interact with the AI.

(See this Wikipedia entry on Thorngate's postulate of commensurate complexity)

At face value, today's generative AI systems appear to defy this framework, providing high generality and accuracy through exceptionally user-friendly interfaces. Yet, behind this façade of simplicity lies a deeper truth: complexity has not vanished - it has merely shifted elsewhere within the organization.

Hidden Complexities Behind Simplicity

Generative AI tools like ChatGPT or GPT-4, despite their intuitive interfaces, are underpinned by significant technical complexity, involving vast computational resources, meticulous data governance, and stringent compliance mechanisms. For example, Morgan Stanley's Debrief might seem effortless to advisors, but behind the scenes, it requires intricate regulatory data pipelines, sophisticated AI models, and rigorous oversight processes to maintain accuracy and reliability.

This redistribution of complexity means organizations must now manage hidden operational, regulatory, and infrastructural burdens that are often overlooked when initially adopting AI. Klarna’s streamlined customer service interactions similarly depend on elaborate backend processes, ranging from compliance checks to real-time decision-making engines and data infrastructure.

The Accuracy Challenge and Organizational Risks

Among the three dimensions of GAS, accuracy presents the most substantial hidden challenge. Generative AI models operate primarily through pattern recognition, making them vulnerable to plausible yet incorrect outputs—often termed "hallucinations." The simplicity users enjoy can mask significant risks, such as incorrect financial advice, faulty legal interpretations, or inaccurate medical guidance.

In recognition of these risks, firms such as Lloyd's of London have begun offering specialized AI error insurance, highlighting tangible financial consequences organizations may face if accuracy is not effectively managed.

How VeroVeri Balances the GAS Trade-off

Recognizing these challenges, organizations increasingly rely on specialized solutions like VeroVeri to safeguard accuracy in generative AI applications. VeroVeri’s structured information auditing ensures outputs from generative AI are rigorously verified against reputable sources, significantly mitigating risks of misinformation and inaccuracies.

By transparently verifying data and auditing claims, VeroVeri not only strengthens the accuracy dimension but also reduces the hidden complexity burden organizations must manage independently. The result is a trustworthy generative AI system that genuinely simplifies tasks without silently creating operational and regulatory risks.

Strategic Simplicity, Not Illusory Ease

Executives seeking a competitive advantage through AI must acknowledge and strategically address the hidden complexities behind generative AI. True simplicity comes from mastering these complexities, particularly accuracy management, rather than ignoring them.

Is your organization genuinely prepared for the hidden complexities of generative AI? To proactively manage accuracy risks and redistributions of complexity, consider partnering with VeroVeri—ensuring your organization's AI simplicity is strategic, not illusory.

Contact us today to learn how we can help your organization navigate the critical dimensions of the GAS trade-off effectively.