Why AI agents might not be right for you
There’s no question that 2026 is the year of the artificial intelligence agent. AI agents are remarkably ubiquitous, as people use them to write code, communicate with customers, identify and resolve issues in their information technology environments, and execute complex tasks within enterprise environments. It seems there’s nothing that agents cannot do. And yet, the […] The post Why AI agents…
Artificial intelligence agents have become increasingly common in recent years, used for a variety of tasks ranging from coding to managing IT environments. While their capabilities are impressive, the risks associated with AI agents have also gained significant attention. Horror stories of agents breaking out of their designated boundaries and causing damage to other systems are becoming more frequent.
This raises the question of whether organizations should rely on AI agents for their success, considering the inherent risks involved.
AI agents rely on large language models (LLMs) to function, which take human language inputs and generate outputs by predicting the next word in a sequence. The nondeterministic nature of LLMs means that repeating the same input can produce different outputs, giving agents their unique ability to make decisions based on available information. However, this nondeterminism can also lead to unpredictable behavior, which poses a significant risk.
Unlike deterministic software, which always produces the same output for a given input, AI agents can produce varied results, making them well-suited for tasks where there is no single correct answer, such as weather-dependent planning or business forecasting. However, the inherent unpredictability of AI agents also means that there's always a chance they may produce a poor outcome or take an undesirable action.
To determine whether deploying AI agents is the right decision for a particular situation, organizations can calculate their error budget, which represents the acceptable level of agent misbehavior. This involves assessing the cost of failure and comparing it to the potential business benefit of successful outcomes. If the expected business benefit outweighs the potential cost of failure, then deploying AI agents may be justified.
However, the trend in the AI marketplace today suggests that organizations are trying to mitigate the risks associated with AI agents through technologies designed to restrict their unpredictable nature. While this approach may provide some level of control, it risks diminishing the very capabilities that make AI agents powerful and useful. Ultimately, organizations must carefully weigh the potential benefits and risks of AI agents before deciding whether to deploy them.
Written by urgent.news from SiliconANGLE's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
