Deep Dive Brief

  • TeamViewer reported on Wednesday that employees' comfort and confidence in using AI are lower than those of their managers, and the power dynamics between AI decision-makers and employees forced to use the technology may exacerbate this trend. The remote access software company, in partnership with Sapio Research, surveyed 4,200 managers and employees globally this spring.
  • Although three-quarters of respondents, including both employees and managers, said they use AI at least once a day, 61% prefer that AI not have independent action capabilities. More than half of respondents frequently or always verify AI outputs, while about the same proportion said they cannot always judge when to trust AI and when to verify it.
  • TeamViewer CEO Oliver Steil said in a press release that companies must commit to achieving trusted autonomy of AI systems, rather than full automation. He said, "AI is already changing how people work, but the next phase will be defined by trust."

Deep Insights

Although both managers and employees are becoming more comfortable using AI at work, leaders driving adoption typically feel better about AI than frontline employees and are more willing to delegate tasks to AI. Employees report feeling greater personal risk from autonomous systems due to a lack of control over how and when AI is deployed.

TeamViewer Chief Product and Technology Officer Mei Dent said in a press release that software has traditionally been designed around tasks that people instruct it to perform, but AI systems and agents change this model.

Insecurity about AI outputs has not led employees to completely avoid the technology, but it may limit its scope of use. The survey found that AI is most commonly used once or twice a day, typically to help users understand information, generate starting points, or complete routine tasks—outputs that can be reviewed before use.

Only 11% of respondents said they are comfortable with AI working autonomously, meaning taking action without review and approval.

Despite lagging trust in AI systems, companies are still advancing their autonomy. A survey released by Kyndryl in June showed that a quarter of business and technology leaders fully trust their AI systems, but more than four-fifths expect autonomous AI agents to make decisions with significant business impact within the next year.

Steil said that organizations struggling with trust often lack the governance and accountability frameworks needed for AI outputs. Respondents worry that AI systems may leak sensitive information, may not understand the context of information to generate correct outputs, and that AI may act beyond its intended scope. About one-fifth of respondents worry they will be held responsible for AI errors.

Steil said, "Organizations will succeed by building systems that operate safely, transparently, and responsibly within clear guardrails. The real opportunity is to use AI and automation to prevent disruptions, improve digital experiences, and allow people to focus on higher-value work."

Steil said organizations must reassure employees that autonomous systems are safe, explainable, reversible, and governed by humans.

A good starting point is to determine which actions are low-risk and can be pre-approved for AI to perform autonomously, and which require human review. Employees should be involved in determining autonomy to understand the permissions systems have and where responsibility for errors lies.

Organizations should also strive to ensure that the risks of autonomous systems are fairly distributed across different work groups, rather than borne by a single team or individual.

Dent said, "Autonomy will scale when organizations can prove it works in specific trusted use cases, and then they can expand with confidence."