Image: © Chanelle M/peopleimages.com/Stock.adobe.com
Dr Noel Carroll, from the University of Galway, provides an alternative view of the much-hyped advances in AI, and explains how low-code/no-code development can benefit from educated use of AI technologies.
Since the time of the ancient Greek philosophers and rhetoricians, there has been curiosity and debate about human intelligence and reasoning, about humans make decisions and arguments.
Artificial intelligence (AI) is one of the most significant technological advancements of our time, with the potential to transform the way we live, work and interact with the world around us.
There has been significant hype around developments such as ChatGPT, ChatSonic and Google Bard AI, which were greeted with a mix of curiosity, excitement, fear and anxiety.
Despite its many benefits, AI is not a silver bullet and has its limitations. It is important to bring a little balance to the hype, and to highlight some of the limitations and to emphasise the importance of human capabilities to co-create value from AI-enabled innovations.
Limits of AI
One of the most significant limitations of AI is its inability to replicate human intuition and creativity. While AI can analyse vast amounts of data, provide insights or construct images based on a set of instructions, it lacks the ability to make judgments based on intuition and experience, which is an essential aspect of human decision-making. This is particularly evident in domains such as art, music, and writing, where AI can produce impressive outputs but lacks the depth and creativity that humans can bring to these fields.
AI algorithms require vast amounts of data to function effectively, and the quality of the data can significantly impact the accuracy of AI systems. Incomplete or biased data can lead to incorrect conclusions, and AI systems may struggle to identify patterns or relationships in the data if it is not appropriately labelled or organised.
Additionally, AI algorithms can be easily fooled by adversarial examples, which are designed to trick the system by making small, subtle changes to the input data. We also know that there are growing concerns around algorithm bias or AI bias, which refers to the tendency of algorithms to reflect human biases.
AI is also limited by its inability to explain its decisions or reasoning. Many AI algorithms are based on deep learning, which involves training a neural network on vast amounts of data. While this approach can be effective, it is challenging to understand how the AI system arrived at a particular conclusion or recommendation. This lack of transparency can be a significant concern, particularly in applications such as healthcare, where the decisions made by an AI system can have life-or-death consequences.
‘AI can produce impressive outputs but lacks the depth and creativity that humans can bring to these fields’
AI cannot understand the context and meaning behind data. For example, an AI system may be able to recognise the words in a sentence but may struggle to understand the nuance or sarcasm behind the text. This can lead to misinterpretations and errors in applications such as natural language processing (NPL) and sentiment analysis.
It is worth noting that ChatGPT is a large language model (LLM) which is built with a Transformer Model Architecture. A transformer is a deep learning model that adopts the mechanism of self-attention, differentially weighting the significance of each part of the input data.
AI is limited by the complexity of the tasks it can perform. While AI has made significant strides in recent years, it is still limited in its ability to perform complex tasks that require higher-level reasoning and decision-making.
For example, while an AI system can identify objects in an image, it may struggle to understand the context or significance of those objects in the broader scene.
Finally, AI is limited by its lack of common sense. For example, an AI system may be able to recognise an object in an image but may not understand the purpose or function of that object. Similarly, an AI system may be able to translate text from one language to another but may struggle to understand the cultural nuances and idioms that can be challenging for non-native speakers to grasp.
Challenge of recruiting talent in AI
A key challenge for organisations trying to keep up with the pace and scale of AI innovations is with the recruitment of talent. Recruiters face many obstacles, including:
- There is a high demand for skilled AI professionals, and the supply of talent has not yet caught up.
- AI is a complex field that requires advanced technical skills, including expertise in machine learning, deep learning, natural language processing, and data analytics. Finding candidates with the necessary technical skills can be challenging.
- AI professionals are in high demand and can be lured away by other companies offering more attractive compensation packages or opportunities for professional growth.
- The AI industry lacks diversity, and there is a shortage of women and people from underrepresented groups in AI roles.
- The education and training required to become an AI professional is often expensive and time-consuming. This can make it difficult for people from disadvantaged backgrounds to enter the field, exacerbating the diversity problem.
To address these challenges, companies may need to re-evaluate their hiring practices to ensure that they are attracting a diverse range of candidates and offering competitive compensation packages and opportunities for professional growth.
Citizen development
In some cases, organisations are turning to their existing employees to upskill and availing of new business trends such as ‘citizen development’, a new method of delivering low-code/no-code (LC/NC).
Citizen development hides the sophistication and complexity of coding but empowers subject matter experts to design, develop and deploy applications into production as though they were full-on, experienced coders.
The trend toward adoption of citizen development is being driven by growing investments in LC/NC platforms by both established and start-up suppliers.
LC/NC development platforms use visual interfaces and drag-and-drop tools to simplify the application development process, allowing users to create and customise applications with little to no coding.
LC/NC reduces the barriers to entry for people and empowers them to play a role while improving inclusive measures around data and digital literacy.
AI technologies can be used to automate many aspects of software development, from generating code to testing and debugging applications.
By combining AI and LC/NC development, businesses and individuals can build applications faster, with greater accuracy and with fewer errors. AI can also help to improve the quality of LC/NC applications, by identifying and fixing potential issues before they become problems.
The potential applications of AI and LC/NC development are vast, including in areas such as customer service, finance, healthcare and logistics.
‘AI has tremendous potential, but it is not without its limitations’
For example, AI-powered chatbots can be developed using LC/NC platforms to provide instant customer service to website visitors. Similarly, LC/NC development platforms can be used to create customised financial reporting tools or healthcare applications that can help doctors and patients manage health conditions more effectively and manage patient flow more efficiently.
Having identified this growing trend and through collaboration with the Project Management Institute (PMI) and Microsoft, my colleagues and I have introduced the Citizen Development University Hub.
The University Hub curriculum centres on leveraging LC/NC platforms to expedite digital transformation.
Student feedback on citizen development and digital, data, and AI-related topics has been extremely positive and students with non-technical background have developed skills, competencies and confidence in these topics.
We have also carried out research to examine how companies such as Shell are adopting and scaling citizen development for digital, data and AI solutions. Companies are beginning to address many of the skill shortage issues through the adoption of citizen development.
AI has tremendous potential, but it is not without its limitations. While these limitations are not insurmountable, they do highlight the need for continued education, research and development to ensure that AI systems can be used effectively and responsibly in a range of applications. It also highlights the importance of critical thinking, curiosity and creativity.
By Noel Carroll
Noel Carroll is an associate professor in Business Information Systems at the University of Galway and founder of the Citizen Development Lab.
10 things you need to know direct to your inbox every weekday. Sign up for the Daily Brief, Silicon Republic’s digest of essential sci-tech news.
