Industry executives and experts share their predictions for 2024.  Read them in this 16th annual VMblog.com series exclusive.

5 Generative AI Predictions

By Greg Benson, Professor of Computer Science at the University of San Francisco and Chief Scientist at SnapLogic

This year, the world was taken by storm with the explosion of ChatGPT and related generative AI technologies. We've witnessed the technology accomplish tasks such as synthesizing existing information, debugging code, and generating data pipelines in a much more efficient way than what is possible through web searching and manually collecting information. And we're only at the tip of the iceberg of witnessing how this technology will revolutionize the industry and impact modern organizations.

Here are the Generative AI predictions that I believe will come to the forefront of the industry in the coming year.

Prediction 1 - GenAI won't take your job, but it might change it.

In 2024, GenAI won't lead to mass job displacements and redundancies, as many early sensationalist reports might have suggested. In fact, GenAI won't completely replace experts in any field, as although models have access to an insurmountable amount of information, users still have to articulate concepts well enough to get the right answers - thus, expertise and human input and, more importantly, human review always will be necessary.

Collaboration with GenAI is a trend we can expect to continue in 2024 as businesses look to capitalize even further on its benefits, reaping the rewards of increased productivity and quality of content creation. This means adopting even more GenAI tools and encouraging even more use with the goal of not replacing workers, but instead assisting what they do.

Prediction 2 - Now that we've invented GenAI, the next step is understanding it.

Next year, we can expect to see businesses attempt to improve the consistency of output from GenAI. Currently, there is no set rule book for achieving great results with GenAI; there are tips and tricks you can deploy for better or faster results, of course, but overall the process is largely trial and error.

Interacting with GenAI in its current iteration is like a science experiment - you come up with a hypothesis and continue to test different manners of prompts until it produces the result you're looking for. In the future, the focus of experimentation will be on figuring out how we evaluate the responses it gives us and using that data to inform prompts further. We will see companies providing tools and services to help run such experiments.

Companies that want to apply GenAI to their products will need to think about how to rigorously assess generative quality and put systems in place to monitor and continuously improve their generative results.

Prediction 3 - GenAI regulation is essential to adoption.

Regulating GenAI will be a huge focus for governing bodies and business leaders in 2024. The EU Parliament has taken the lead with the provisional agreement on the Artificial Intelligence ACT. This needs to be fleshed out further, but other governments will take notice and develop their own AI regulations.

Earlier this year, calls were heard from numerous visionaries for a pause in AI development, but this isn't realistic as the fundamental technology is increasingly available through open-source models on Hugging Face and elsewhere. Rather than focusing on halting development, creating clear regulations, guidelines, and best-use practices will be necessary to ensure that partnership with AI moves forward in a safe and secure way.

Like any other technology, defining the boundaries that keep safety in mind will allow for leveraging the benefits without sacrificing progress. We can liken this to all manners of tools and equipment that need to be regulated; for example, we don't stop ourselves from building cars that go really fast, but we do put speed limits in place to ensure safety. Internationally, governments will draw their attention first to the areas of regulation that present the greatest impact on citizens, including frontier AI.

From an industry perspective, the GenAI applications and use cases that are most helpful will emerge as front runners for wider business use cases. Understanding the risks, challenges, and security issues potentially imposed by these tools will be vital for businesses to understand exactly when and how these tools need to be regulated internally. Likewise, companies hoping to leverage GenAI will have to communicate to customers exactly how it's used and how it complies with current and future regulation requirements.

Prediction 4 - GenAI and Legacy technology: Why the key to modernization may reside in GenAI tools.

After a year of GenAI practice, legacy businesses are starting to understand that GenAI interest is not just driven by ‘hype' and instead could be truly transformative for their sector. Therefore, in 2024, we can expect even more traditional businesses to deploy the technology to help evolve legacy systems and modernize their technology stack.

Typically, traditional companies are not amenable to change or agile enough to adopt the latest in new technology. Many companies are tied to legacy software due to a combination of outdated procurement processes, familiarity, or concerns about data loss or disruption, making modernization inaccessible. The key here is that GenAI can assist with migrating old code bases and technology stacks to modern programming languages and platforms. There is great potential for such automated migration to help companies reduce costs by moving off legacy systems.

Prediction 5 - Universities will begin to teach prompt engineering

In 2024, universities will teach prompt engineering as a minor field of study and through certificate programs. While GenAI has created a bit of a firestorm in higher education in that students can get answers to their homework problems, it is also an opportunity for colleges and universities to help shape how students engage with the technology to use it productively and responsibly.

Prompt engineering for GenAI is a skill already augmenting domain experts, similar to how computing has augmented other domains. The successful use of large language models (LLMs) relies heavily on giving the models the right prompts. When looking to fill the role of a prompt engineer, the task becomes finding a domain expert who can formulate a question with examples in a specific domain, a skill critical for today's IT professionals to refine to successfully implement LLM applications. Given this, universities will introduce both general prompt engineering and discipline-specific prompt engineering to address the growing demand for professionals with the skills required to build the next generation of GenAI applications.

##

ABOUT THE AUTHOR

Greg is the Chief Scientist at SnapLogic and leads forward looking research and innovation projects. He is also a Professor of Computer Science at the University of San Francisco. Greg has published several papers in the fields of operating systems, distributed systems, and programming languages. He is one of the original architects of the SnapLogic programming model and cloud platform. Greg also led the machine learning research work that resulted in SnapLogic's Iris AI.