A recent survey sponsored by IBM has shed light on the numerous challenges faced by AI developers, highlighting a significant skills gap, frustration with immature processes, and inadequate tools. The survey, which gathered input from over 1,000 enterprise developers in the US, paints a stark picture of the difficulties developers encounter when building generative AI applications.

One of the primary challenges identified in the survey is the skills gap, with less than one-quarter of application developers (24%) considering themselves experts in generative AI. Even among AI developers and data scientists, only a majority (51% and 55%, respectively) consider themselves experts in the field. This lack of expertise is compounded by the rapid pace of innovation in the generative AI space, making it difficult for developers to keep up with the latest techniques and technologies.

In addition to the skills gap, developers also face challenges related to process and tooling. The survey found that a lack of standardized AI development processes and inadequate tools are major hurdles, with over a third of respondents citing these as top challenges. Furthermore, developers are frustrated with the current state of tools, with many feeling that they do not meet their needs in terms of performance, flexibility, ease of use, and integration.

The survey also explored the use of tools among developers, finding that the majority use between five and 15 tools to do their jobs. However, these tools often lack essential qualities such as good documentation, cost-effectiveness, community support, and open-source availability. As a result, developers are reluctant to invest time in learning new tools, with two-thirds willing to spend only two hours or less on learning a new AI development tool.

Another area of concern for developers is the development of AI agents, with 99% of respondents exploring or developing AI agents. However, this comes with concerns around trustworthiness, introducing new attack vectors, compliance, and losing oversight and visibility into systems. Despite these concerns, the top use cases for AI agents are customer service and support, project management/personal assistant, and content creation.

The survey's findings highlight the need for improved tools and techniques to support AI development. According to Ritika Gunnar, general manager of Data and AI at IBM, "the AI development stack doesn't receive a lot of attention in the broader generative AI conversation. Yet it can play an outsized role in the technology's impact." Gunnar emphasizes the need for tools that are easy to master and can help address the complexity of AI development.

The survey's results serve as a call to arms for the industry to focus on improving the AI development stack, making it simpler and more intuitive for developers to build generative AI applications. By addressing these challenges, developers can unlock the full potential of AI and drive innovation in the field.

In conclusion, the survey provides a stark reminder of the challenges faced by AI developers and the need for the industry to come together to address these issues. By acknowledging these challenges and working towards solutions, we can unlock the true potential of AI and drive meaningful innovation in the field.