Predictions for GenAI adoption in 2024
Generative artificial intelligence (GenAI) entered the public consciousness and debate about one year ago. As a science, it goes back several years but as an applicable piece of software, it is very much in its infancy.
Text, images, and audio can be generated by GenAI models,but their integration into existing software tools worldwide is still in its early stages. Similarly, the majority of business leaders are only talking about GenAI, with some experimenting through proofs of concepts, while a small minority have deployed initial, and usually specific, use cases.
In public and private discussions, most time is spent on the following questions:
- Is GenAI good enough given it hallucinates?
- What is the value of using it in real business life?
- Is it even legal and who is responsible for any mistakes?
- What if it is sentient and becomes a threat to humanity?
AI’s current state of play
In short, the current school of thought is that GenAI is good enough for some applications but not others due to making factual and logical errors called ‘hallucinations’. Users must carefully examine use cases on their own merits. For example, some have reported that the value of the technology lies in an increase of desk worker productivity rates by about 30-40 percent.
From a legal standpoint, given the fact that several lawsuits and governmental regulations are still in flight, the legal status of GenAI is uncertain. However, this will continue to be a complex road, and some years will pass before stable legislation will be available in most major countries.
And lastly, in positive news for humankind, GenAI is very far from sentient. No one is really working on making it conscious because the industry is still deciphering how to make money from what it has — GenAI will certainly not be the end of us.
GenAI in practice in 2024
Delving into the major uses of GenAI, we can expect to see significant adoption in 2024 alongside large measurable business value.
One: Source code generation. Computer programmers, both individually and in teams, have achieved great productivity gains by using GenAI to make first drafts for source code, comments, and documentation. This is the only well-established wide-spread business use case to date that has reliable benefit metrics.
Two: Enterprise search. GenAI is able to search through the entire storage of proprietary information of your company including audio or video recordings. It will generate not a hit list of links but fully formed answers to your questions. Imagine being able to ask questions of thousands of recorded customer phone calls. This is a capability that only exists due to GenAI, and is currently valued at a staggering $4.5b.
Three: Summarisation. Picture being faced with long texts or videos while only needing the gist of what is being covered. GenAI is able to provide a good summary of the content instantly, saving time whilst preserving the essential information.
Four: Customer support. Support hotlines currently work either in a rigid decision-tree mode that frustrates everyone or in human mode where a person attends the customer. GenAI combined with voice capabilities can provide basic support to many simple queries without involving a human agent - and without frustrating the customer. This is another net new capability.
Five: Marketing messages. Creating marketing messages addressed to a single individual sent to everyone on the distribution list was impossible until GenAI. We can now feed GenAI precise content about a customer’s behavior and generate a custom message with much higher relevance than a general blast.
Six: General-purpose interface. The ability to use written or spoken language as a general-purpose interface to software without needing to know precisely where to click will aid in automating processes and workflows.
Seven: Education. Both education for children and life-long learning for adults will see marked changes with GenAI. While more of a personal than a business benefit, we will all be able to learn more effectively with adaptive and active learning technologies, rather than passively consuming static content.
Looking to the future
It is important to remember that we are firmly in the realm of narrow AI, where AI is capable of performing one specific task in a narrowly defined realm very well. GenAI applications can do one task well, and not others. GenAI is also not capable of logical, causal, or common sense reasoning, and should not be asked to make judgment calls or recommendations.
The above use cases showcase genuine, measurable business value for brick-and-mortar companies interested in using GenAI to future-proof their business. These are the uses that I forecast GenAI to be put to in 2024 as the conversation turns into adoption.
Photo Credit: Andrey_Popov/Shutterstock
Patrick Bangert, Senior Vice President for data, analytics, and AI at Searce which provides professional services for cloud applications. He heads the profit center that is responsible for all projects with a data scientific character globally. Until recently, Patrick led the AI Division of Samsung SDS from 2020 to 2023 bringing AI tools and services into Samsung Cloud for computer vision, natural language processing, and machine learning with a particular focus on medical imaging. He holds a PhD in AI and has over 20 years of business experience.