Despite the impressive growth of generative AI technologies like ChatGPT, many companies are still cautious about their use. For example, the tobacco company Reynolds American is only testing generative AI on a limited scale to improve data analysis but avoids it in high-risk financial applications or in direct customer contact. "There are some inherent risks with generative AI," explains CIO Aaron Gwinner. "Before we just launch and conduct AI projects, we need to build the foundations and basic knowledge."

Despite hopes that technologies such as Microsoft's Copilot, Google's Gemini, and OpenAI's ChatGPT could transform the business world, many companies, including Reynolds, are proceeding with caution. Data from the US Census Bureau from March shows that only about 5.4% of companies use AI in some form for the production of goods or services. Above all, large companies and those from the information sector make stronger use of AI.

Companies from the Life Sciences, Oil and Gas, and Communication Services Sectors Express Intent to Adopt Generative AI, Yet Often Fail to Implement These Plans. "Organizations are still trying to understand how useful generative AI really is," says Daniel Colson from the AI Policy Institute.

Some companies doubt the value of technology or can't find suitable use cases. The law firm Luminos Law, which specializes in risk management consulting in the field of AI, has to date not implemented any use of generative AI. "We are small and quite elite, and it just doesn't make sense," says co-founder Andrew Burt.

Koko, an Online Mental Health Service, Attracted Criticism after it was Revealed that ChatGPT had Written Empathetic Messages to Users. Although the Messages were Approved by a Human, Users Found the Artificial Empathy to be Strange and Hollow.

Even Larger Companies that Have Already Adopted Generative AI See Potential for Security Risks. A Survey by Cisco Systems Among Privacy and Security Experts Revealed that 92% of Respondents Believe Generative AI is Fundamentally Different from Other Technologies and Requires New Techniques for Data and Risk Management. More Than a Quarter Even Prohibited its Use.

Time Could Be the Main Factor Preventing Widespread Adoption. Search Engines and Social Media Took Years to Gain General Acceptance. ChatGPT Reached 100 Million Monthly Users Within Two Months of Its Launch in November 2022, Becoming the Fastest-Growing App in History at That Time.

Generative AI May Ultimately Follow the Same Path to Ubiquitous Use as These Prior Technologies. Microsoft Found That 77% of Users Who Tried Their Generative AI Product Copilot Do Not Want to Give It Up.

Reynolds American tries to weigh the competitive advantages of quick action against the risk of moving too fast. "When you think about the revolutions in technology over the last 100 years, this will be one of the biggest," says Gwinner. "[But] it will take time until we are ready for it."