Mass Phenomenon: “Bring Your Own AI”
A recent study by Software AG reveals that 54% of German knowledge workers are using unsanctioned AI tools, a practice known as “shadow AI.” Nearly as many workers would bypass any bans imposed on these tools.
This translates to over half of all employees relying on AI tools not officially provided by their employers. Shadow AI has become so indispensable that nearly half of all workers (46% globally, and 49% in Germany) would continue using it even if their companies enacted a complete ban.
The study surveyed 6,000 knowledge workers from the U.S., U.K., and Germany about their AI habits.
“While 75% of knowledge workers are currently using AI, this number is expected to rise to 90% soon, as it saves time (83%), eases workload (81%), and boosts productivity (71%),” explains Steve Ponting, Director at Software AG. He cautions, “As usage increases, so does the risk of cyberattacks, data loss, and regulatory non-compliance. Organisations need to prepare for this trend before it’s too late.”
Fields of Application
The study also highlights differences in AI use between German workers and their international peers. More than half of German workers (51%) report no “AI FOMO” (fear of missing out), believing they aren’t missing out if they don’t use AI tools—higher than the global average of 46%. German workers are also more likely to use AI for text creation (49%) compared to just 19% who use it for research, while globally, 35% of respondents use AI for research.
Many knowledge workers choose their own AI tools to maintain independence (53%), while 33% say this is due to a lack of suitable tools provided by their IT teams. “This suggests that companies need to rethink how they determine which tools to offer if they want employees to adopt official solutions,” says Marc Vietor, General Manager of Business Transformation at Software AG.
Security Awareness
Most employees are aware of the risks tied to their AI use, including cybersecurity concerns (72%), data governance issues (70%), and the potential for inaccurate information. However, the study indicates that few take precautionary steps, such as conducting security scans (27%) or reviewing data usage policies (29%).
German workers show a higher level of security awareness than their international counterparts. While only 28% of global respondents anonymise sensitive data before inputting it into an AI tool, 34% of German workers take this step.
“Shadow AI compounds operational chaos that is already present in many organisations,” says Vietor. “Companies should invest more in training, as many have yet to implement sufficient safeguards.” He emphasizes the importance of risk management training for occasional users who may be less experienced but equally prone to risky behavior. “Improving transparency and offering targeted training on shadow AI can help mitigate operational chaos.”
Study Methodology
The study, conducted by TEAM LEWIS and commissioned by Software AG, surveyed 6,000 knowledge workers between September 13 and 25, 2024. Respondents included 2,000 participants each from the U.S., U.K., and Germany. All participants, aged 18 and older, were recruited independently by RepData, and the sample was age-balanced according to Software AG’s specifications.
