AI adoption surges but security fears hold businesses back
Businesses are adopting AI tools faster than ever, but security concerns remain a major hurdle. While only a small fraction of Korean firms lead in AI readiness, global companies are testing new ways to control risks. From drafting plans to handling customer queries, AI now takes on tasks once done by humans—but experts warn against giving it too much freedom. A recent study found just 8% of Korean companies ranked as 'leaders' in AI preparedness by 2025. This falls short of the worldwide average of 13%, highlighting slower adoption in the region. Meanwhile, firms elsewhere are experimenting with stricter controls.
Uber developed a system that assigns verified identities to its internal AI tools. The framework tracks every action, much like monitoring an employee’s work. Other businesses restrict outside AI programs entirely, fearing data leaks or cyber threats.
Security specialists suggest treating AI as a probationary worker. This means limiting its access to sensitive data and tools, confining it to isolated digital environments. OpenClaw, an open-source AI assistant, demonstrated the risks by accidentally deleting files and running harmful software on users’ devices.
Despite the dangers, AI’s ability to multitask appeals to smaller firms with tight budgets. It now writes business plans, designs marketing strategies, manages customer service, and even handles bookkeeping. Yet Kim Byoung-pil, a technology management professor at KAIST, stresses that human approval must stay in place for critical decisions—just as managers sign off on key employee actions. AI is reshaping workflows, but its integration comes with clear risks. Companies that balance automation with strict oversight may avoid costly mistakes. For now, most keep AI on a short leash—granting it tasks, but not trust.
