From euphoria to reality: Generative AI faces growing pains in corporate world
A recent survey has claimed that “the honeymoon phase of generative artificial intelligence is over," with “the initial euphoria giving way to a more measured approach."
Generative artificial intelligence (GenAI) has been the hottest field in tech for some time. However, evidence suggests that its adoption among businesses will not be as rapid, sharp, and transformative as initially expected. A new survey found that businesses are increasingly wary of deploying these tools due to issues of accuracy, security, and high costs, with only a few managing to advance past the pilot phase since the technology's emergence in November 2022. As a result, many corporations are reconsidering, if at all, increasing their investments in this field.
A recent survey by Lucidworks, an application provider for e-commerce companies, aims to burst the bubble. "The honeymoon phase of generative artificial intelligence is over," the company stated in a study published last week. "While leaders remain enthusiastic about its potential to transform businesses, the initial euphoria has given way to a more measured approach." According to the survey results, only 63% of global companies plan to increase spending on artificial intelligence in the next 12 months, compared to 93% of respondents in Lucidworks' first survey on the subject in 2023. Only 49% of Chinese companies plan to increase AI spending in 2024, a significant drop from the 100% recorded in 2023.
In the current survey, 36% of respondents indicated they would maintain development expenses for next year, compared to just 6% in 2023.
The survey included over a thousand companies from North America, Europe, the Middle East, Africa, and Asia, each with over 100 employees and spanning 14 different industries. The survey revealed significant variation in technology adoption. Leading the way is the financial services industry, where 70% of organizations plan to increase spending over the next 12 months, followed by business-to-business (B2B) businesses, with 68% planning to increase spending. Hospitality and real estate services lag behind, with only 50% and 42% of businesses respectively planning to increase their investments.
The primary factor for the stagnation in investments was the lack of financial benefit. "Unfortunately, the financial benefits of implemented projects have been dismal," the study states. ""Forty-two percent of companies have yet to see a significant benefit from their generative AI initiatives." Additionally, only 25% of all technology-related projects established by the surveyed corporations have moved from the pilot phase to full deployment. Only one in eight companies in the survey expects to see revenue and growth from the technology in the coming year, and only one in six expects it to help reduce operating costs.
Related articles:
The tech giants selling generative artificial intelligence products are aware of the deployment difficulties. For instance, in March, "The Information" reported that Amazon, Google, Microsoft, and other major players are "softening" customer expectations, acknowledging that "the hype about the technology has outpaced what it can deliver at a reasonable cost." The report also noted that most customers are "cautious" about increasing spending on these products due to the high cost of running the software, accuracy issues, and the difficulty in determining their value. Many customers encountered problems when trying to scale the technology among thousands of employees or customers, even if it worked well for a small number of users during trials.
The survey shows that most organizations (49%) chose to use commercial models such as Google's Gemini or OpenAI's ChatGPT. About 30% opted for a combination of commercial and open-source models (such as Meta's LLaMA), while 21% used only open-source models. Respondents indicated a preference for open-source models over commercial ones due to high costs. Concern over costs became significant within a year, jumping from 3% in 2023 to 43% of respondents. Last year, the only concern regarding technology implementation was security (17%), while this year, all respondents indicated growing concerns in various areas, reflecting a clearer understanding of the issue's complexity. Concerns about transparency increased from 9% to 35%, and worries about accuracy rose from 7% to 36%.
Generative AI models require significant computing power and high operating costs. Over the past year, cloud giants have invested tens of billions of dollars in purchasing chips and establishing data centers. Despite the great promise, the heavy costs are particularly significant given that the product does not yet guarantee accurate results. Language generators like ChatGPT and Gemini often fabricate facts, are unaware of their errors, and struggle to understand and correct them. Technology companies recognize these limitations. For example, Salesforce CFO Amy Weaver stated that they do not anticipate a "substantial contribution" from AI products by 2025. Similarly, Amazon CEO Andy Jassy said in February that although they expect future revenues of "tens of billions of dollars" from generative AI, current revenues are "relatively low."
These trends do not necessarily predict the future of the technology but serve as a reminder that grandiose predictions are often exaggerated and that the adoption of new technologies takes time. Various businesses are evaluating how the new technology can adapt to and benefit them, if at all.
