Background 

Among the more recognizable impacts of software engineering in the recent past has been the rise and use of artificial intelligence to solve real-world business problems. Artificial intelligence is starting to acquire catalytic importance in a company’s digital foundation as complexities multiply in the technology and business domains. The year 2023 could see significant strides in the use of artificial intelligence to bolster digital transformation initiatives. 

Democratized AI – Leading the trend in 2023 

Current trends suggest that AI will get generally democratized. Additionally, we will observe an increasing application of composite AI use cases. Digital trends that are emerging indicate that they will revolve around resilient, robust, and trusted operations that can quickly scale horizontally or vertically. Enhanced digital engagement and accelerated time to market to drive more opportunity will be built on the back of digital transformation initiatives, as current trends going into 2023 suggest. 

There is a lot to unpack here, but this article will explore the democratization of AI briefly juxtaposed with the advancement of composite AI to solve business problems. Given the growing mounds of data and data sources that can be securely accessed, analyzed, and learned from, the number of business problems that can be solved with AI has been growing by leaps and bounds. It has resulted in a more democratic world of AI. Now, what does one mean by Democratized AI? A simple definition states that democratized AI means AI applications that more and more everyday users use. It follows then that democratized AI would include a set of AI tools and techniques that lessens the burden on humans and reduces the need for expert knowledge. Organizations can now eschew complex data science roles and AI expertise with simplified AI solutions that do a lot of this grunt work. A classic case is having AI solutions do what expensive data scientists and analysts would have traditionally done. Instead of manually scouring through reams of data to decipher patterns and other behavior, AI solutions can instantly provide this information to an everyday business user.

Multiple AI approaches continue to be used depending on the needs of the use case. Specific characteristics of business use cases can significantly influence the end goal of the AI solution. Organizations will consider several parameters to decide on AI approaches depending on where they are in their digital transformation journey. In most cases, it simply boils down to where they are on the decision-risk to task-complexity matrix. In 2023, organizations will start looking at AI approaches through the lens of the decision-risk to task-complexity matrix, shown below. 

Fig 1- Automate or Augment using AI? 

As we look at the figure above, we see that organizations will invest in fully automated AI solutions where the task complexity is low and the decision risk is low. As we progress to higher decision risks and higher complexity tasks, we observe that augmented AI will be preferred to support a business user. Some tasks are so complex that a human in the loop will be required, and AI solutions become aids for decision support. The science of using the right approach is still in its infancy and will evolve as more sophisticated use cases come into play and we learn more from its practical uses. 

2023 Sub-trends 

2023 will see some sub-trends of democratized AI emerge more robust than others. 

• The most common sub-trend is the move towards fully automated Artificial Intelligence solutions. As AI techniques begin to be embedded within everyday applications, we will see fully automated tasks like scheduling and notifications and more complex applications with lower decision risks. 
• Another sub-trend becoming very popular is the role of augmented AI. As intelligent process applications start gaining popularity, we will see an increasing trend to use augmented AI. For example, enterprise business applications with embedded or integrated AI technologies, such as intelligent automation; data-driven insights and guided recommendations, that can deliver a more personalized interface, improve productivity and support decision making. 
• A new trend gathering a lot of traction is that of no-code AI, where engineers with little to no coding knowledge can quickly train or retrain existing AI models and tune them to make them more effective and relevant. 
• An exciting trend, doffing its hat to the ingenuity of human wisdom and the difference between knowledge learned and wisdom earned, is using a human-in-the-loop approach to AI. It assumes a seamless partnership between humans and machines to complete complex business applications that require advanced cognitive judgement, subjective decision making and unusual pattern recognition. 

2023 Trend – Evolution of Composite AI 

As described above and for reasons stated earlier, democratized AI will dominate digital transformation initiatives into 2023 and beyond. Some of the trends that started in 2022 will continue to get stronger. One trend expected to gather more steam is using various AI techniques to create composite AI applications. Many use cases are not machine trainable, given the need for more data in some fields. Even though the general availability of data is high, companies are wary of putting a lot of data in the public domain for sensitive business applications. In such cases, domain knowledge and human expertise will be required to give context to AI models and require continuous human retraining. Some techniques may entail using knowledge engineering, Auto ML, computer vision, natural language processing or domain-specific language processing like engineering language processing and application of advanced spatial, temporal, or relational intelligence techniques. 

As the year 2023 rolls in...

It will become increasingly evident as the year 2023 dawns that the democratization of AI and the use of multiple AI techniques in composite AI applications will trend. We will see a massive jump in the sophistication and complexity of the AI engineering discipline and see more software engineers and programmers begin to gravitate towards this field. We will see more business problems using AI as their digital foundation to solve them. All in all, a fascinating time to be in the Artificial Intelligence arena. 

Photo by DeepMind on Unsplash