AGENTIC AI
Digital Teammates Salesforce has introduced Agentforce in Slack, a platform that enables the creation of task- specific AI "digital teammates". These agents can access company data, understand Slack conversations, and perform multi-step actions within the Slack environment.
We’ve long accepted that humans are the weakest link in security... ...but AI agents
The digital teammates are designed to go beyond traditional chatbots and offer more collaborative, proactive capabilities. One could say, “become more like us”, as in becoming one, or more, of the employees. However, AI agents struggle with ethics, subtleties, nuance, bias, context, the list goes on, making them especially vulnerable to manipulation and coercion. They’re pretty impervious to a disciplinary action too, other than to discipline them with re-prompt engineering. AI agent deployments are expected to grow 327% during the next two years. Elad Koren, VP of Product Management at Palo Alto Networks wrote in a recent blog, “but from the vantage point of cybersecurity, this evolution introduces a volatile mix of innovation and risk. We’re not just giving software system access — we’re giving identity, autonomy, and decision-making capabilities. That changes how organisations approach security entirely.” Koren forebodes, “Let’s be clear: These AI agents are not tools in the traditional sense. Unlike conventional automation or service accounts, these agents act as authenticated users operating under corporate credentials, making decisions, interacting with systems and data, and in some cases, executing sensitive tasks. That means they will have the same access and arguably pose the same risks as a human employee. But unlike humans, AI agents don’t understand context, intent, or consequences the way we do. They can be tricked, manipulated,
or coerced through techniques like prompt injection or adversarial inputs. We’ve long accepted that humans are the weakest link in security – phishing and social-engineering schemes prey on our psychology – but AI agents introduce an even softer target: They take things at face value, don’t call the help desk, and operate at machine speed. Once compromised, they could serve as a persistent, high-bandwidth attack surface buried deep inside an organisation’s environment.”
introduce an even softer target.
So, as companies roll out these shiny new “digital teammates” to take on tasks and chat away in Slack, maybe it’s time we treat them a bit more like office rookies: give them a solid orientation, keep an eye on their work, and maybe don’t let them hold the keys to the company safe just yet. n
wu cwawd v. naenwc es idn.tchoemc h a n n e l .c o. u k
49 5337
Powered by FlippingBook