ChatGPT - Comparison of the versions

The use of AI in the enterprise is popular. So popular, in fact, that more than half of employees using AI are using unauthorized AI tools. How are organizations dealing with this shadow AI?

The increasing use of AI in the enterprise context has achieved remarkable popularity. According to a Salesforce study, more than half of German employees using generative AI are using unauthorized tools. Even in organizations that officially provide AI tools, one-third use alternative solutions. Globally, 28% of all employees use generative AI, illustrating the scale of the shadow AI phenomenon.

Quelle: Salesforce

How to deal with Shadow AI?

In light of these numbers, organizations are faced with the question of how to manage the covert use of AI. These are the most important aspects of a solution:

1. Allow the use of AI tools: One key is to legalize the use of AI tools. Safe options already exist with products such as ChatGPT Team and Enterprise or Microsoft's Copilot. These applications guarantee that user input and responses are not used to train the speech model. In addition, their vendors protect their customers from copyright infringement claims. Alternatively, many companies rely on customized chatbots, which are more cost-effective and can address specific business needs. For example, by leveraging internal company data, they can simplify the onboarding of new team members or inform marketing strategy with information from past campaigns.
2. Regulate the use of AI tools: The use of AI requires clear guidelines. Even the most secure AI tools should not process sensitive personal information. Every organization needs to develop an AI policy that educates employees on the importance of safe and ethical AI use. These policies define clear boundaries and create a framework in which everyone feels safe to explore AI.
3. Train people to use AI tools: Once the tools and policies are in place, it's time to build employees' AI skills. This includes training on safe and ethical use, as well as learning the best use scenarios for the technology. Because the use of AI in the workplace is not a passing fad, but a permanent change, these training programs should be viewed as an ongoing initiative.
4. Allow for trial and error: A culture of experimentation can further drive adoption and encourage employees who may have previously relied on unauthorized tools to use their skills within company policy. This can not only increase their motivation, but also lead to the development of new, productive use cases.

Successfully managing change requires more than just policies and tools. Rather, a corporate culture based on openness and innovation will create the framework for AI to come out of the shadows and realize its full potential for the business.