Generative AI has dominated the technology landscape in 2023, with popular tools such as ChatGPT, Stable Diffusion, GitHub Copilot, and Midjourney capturing the attention of the masses. Some users have even experimented with Bard, Claude, and LLaMA1. The impact of these language models and art generation programs on work dynamics, the possibility of singularity, and even the potential doom of humanity has sparked intense debates. In the corporate world, we have witnessed both widespread adoption and restrictive policies that ban or limit the use of generative AI. To gain a better understanding of the reality, we conducted a survey of O’Reilly’s users in September. Our survey focused on how companies are utilizing generative AI, the challenges they face in adoption, and the skills gaps that need to be addressed.
Executive Summary:
Generative AI has experienced an unprecedented level of adoption, considering that ChatGPT is barely a year old. As of November 2023, our survey found that 67% of respondents’ companies are using generative AI. The most in-demand skills are AI programming (66%) and data analysis (59%). While 26% of respondents have only been working with AI for under a year, 18% already have applications in production. The main barriers to adoption are the difficulty in finding suitable use cases and the associated risks, including unexpected outcomes, security, safety, fairness, bias, and privacy. However, 54% of AI users anticipate that the biggest benefit of AI will be increased productivity, while only 4% believe it will lead to reduced headcounts. Despite the hype surrounding generative AI, there is still significant room for growth as adopters explore new use cases and reimagine business processes.
Users and Nonusers:
While generative AI adoption is on the rise, it is not yet universal. Our survey revealed that 67% of respondents’ companies are using generative AI, with 41% having utilized AI for over a year and 26% for less than a year. Only 33% reported that their companies are not utilizing AI at all. Comparing generative AI users to nonusers, the majority (38%) of users have been working with AI for less than a year and are still in the early stages of experimentation and proof-of-concept projects. It is important to note that even with the availability of cloud-based foundation models like GPT-4, which eliminate the need for developing custom models or infrastructure, fine-tuning a model for specific use cases remains a significant undertaking. While the rapid pace of adoption suggests that generative AI may be at the peak of the hype cycle, we believe there is still ample room for growth. However, AI must prove its value to new adopters to sustain this momentum and avoid potential setbacks.
What’s Holding AI Back?
To understand the reasons behind the limited adoption of AI, we asked nonusers why their companies were not utilizing AI and users what bottlenecks were hindering further adoption. Both groups were given the same set of answer options to choose from. The most common reason cited by nonusers (31%) and users (22%) was the difficulty in finding appropriate business use cases. This highlights the importance of thoughtful consideration and imagination when applying AI, as haphazard implementation can have negative consequences. The second most common concern for both groups was legal issues, risk, and compliance (18% for nonusers, 20% for users). The legal implications of using generative AI, such as copyright ownership and potential violations, are still uncertain and will likely be resolved through court cases in the future. Other risks include reputational damage from inappropriate output, security vulnerabilities, and more. Additionally, the absence of company-wide policies for AI use (cited by 6.3% of users and 3.9% of nonusers) adds to the challenges. However, this issue is expected to be addressed as corporate policies on AI use evolve. While the lack of policies does not currently hinder AI adoption among users, it is a matter that should be prioritized to mitigate risks and liabilities. Overall, careful consideration of use cases, understanding of AI’s transformative potential for businesses, and the establishment of appropriate policies are crucial for the successful adoption of generative AI.
Source link
