Companies Trust AI’s Advice but Not Its Actions

Source Domain: www.pymnts.com

Companies are increasingly utilizing artificial intelligence (AI) in various operations, yet still largely restrict AI agents to only retrieve information without allowing them to autonomously act on their own.
A survey of 60 senior technology executives in large enterprises in the United States showed that across all sub-industries, 100% of wholesale firms, 90% of retailers, and 85% of construction firms limit AI agents to look-up access only.
While AI is being embedded broadly for research, analysis, and document generation, no companies currently allow AI agents to take autonomous actions.
Although most companies plan to move toward semi-autonomous AI with human oversight within five years, none currently or in the future is aiming for fully autonomous AI action.
The restriction on AI actions is a governance decision due to concerns about risks like agents acting without traceable identity, operating outside approved boundaries, or exposing sensitive data. Trust remains a larger barrier to fully autonomous AI adoption than technological capability.

You may have missed