AI and Citizen Development: Challenges and Opportunities Coexist

Original: https://cli.im/article/detail/2101

Note: The term "Citizen Development" was proposed by the consulting firm Gartner in 2010. It refers to non-professional developers using low-code or no-code platforms to create applications without the support of the IT department, aiming to improve productivity and reduce development costs.
While "Citizen Development" is commonly translated as "公民开发" in China, CaoLiao QR Code believes that Citizen Development is not merely a technology but a work model and standard. It should be translated as "全民开发," meaning that everyone who understands business can become a developer.

The following is a translated viewpoint article from the renowned tech media Silicon Republic, compiled by CaoLiao QR Code.

While the outside world is generally "hyping" artificial intelligence, Dr. Noel Carroll from the National University of Ireland, Galway, offers a different perspective: how no-code developers can benefit from AI technology.

Since the time of ancient Greek philosophers and scientists, debates have centered on human intelligence and reasoning, how humans make decisions, and how arguments are constructed. As one of the most significant technological advancements of our era, artificial intelligence has the potential to reshape our lifestyles, work methods, and interactions with the world.

The emergence of new technologies like ChatGPT, ChatSonic, and Google Bard AI has sparked widespread discussion, evoking curiosity and excitement in some, while instilling fear and anxiety in others. Although AI capabilities at this stage are powerful, they are not omnipotent and have their limitations. It is essential to rationally recognize the capabilities of AI, particularly its shortcomings and constraints, and to deeply understand the importance of human-AI collaborative innovation.

Limitations of AI

One of the biggest shortcomings of current AI is its inability to replicate human intuition and creativity. While AI can analyze vast amounts of data and provide insights or build prototypes based on a set of instructions, it lacks the ability to make judgments based on intuition and experience—a hallmark of human decision-making. This is particularly evident in fields such as art, music, and writing, where AI can generate impressive responses but lacks the depth and creativity inherent to humans.

AI algorithms require large amounts of data to function effectively, and the quality of the data directly impacts the accuracy of AI responses. Incomplete or biased data can lead to erroneous conclusions. Without proper data standards, AI struggles to identify causal relationships within the data.

Additionally, algorithms can be deceived by adversarial examples—subtle adjustments to sample data designed to mislead the system. Concerns about algorithmic bias, or AI bias, are growing, yet the biases exhibited by algorithms often reflect human biases.

AI also faces limitations in explaining its decisions or reasoning. Many AI algorithms are trained using deep learning, which involves training neural networks on massive datasets. While this approach is effective, it is challenging to understand how AI arrives at specific conclusions or recommendations, resulting in a lack of transparency. In critical applications like healthcare, decisions made by AI can have life-or-death consequences.

Currently, AI cannot grasp the context and meaning behind data. For instance, an AI system might recognize words in a sentence but fail to understand the nuances or sarcasm in the text, leading to misunderstandings and errors. This is particularly common in scenarios like natural language processing and sentiment analysis.

It is worth noting that ChatGPT is a large language model based on the Transformer architecture. Transformer is a deep learning model that employs a self-attention mechanism, allowing it to differentially weigh the importance of each part of the input data.

AI is also limited by the complexity of tasks it can perform. Despite significant advancements in recent years, AI's reasoning and decision-making capabilities remain constrained. For example, while an AI system can identify objects in an image, it struggles to understand the significance of those objects within the scene.

Finally, AI is constrained by its lack of common sense. For instance, AI can recognize an image but cannot comprehend its purpose or meaning. It can translate text from one language to another but finds it difficult to grasp cultural nuances, posing challenges for non-native speakers.

Challenges in AI Talent Recruitment

For organizations seeking to keep pace with AI innovation, one of the biggest challenges lies in recruiting AI talent. Recruiting such talent presents numerous challenges, including but not limited to:

• High demand for highly skilled AI professionals, while the supply of talent has not kept up;
• AI is a complex field requiring expertise in machine learning, deep learning, natural language processing, data analysis, and more, making it even more challenging to find suitable candidates;
• The high demand for AI talent means that other companies may offer more attractive compensation and opportunities;
• The AI industry lacks diversity, such as the underrepresentation of women in AI roles;
• Becoming a professional AI talent requires extensive training and education, which may deter individuals from disadvantaged backgrounds, exacerbating diversity issues.

To address these challenges, companies need to reassess their recruitment strategies to attract a diverse pool of candidates and offer competitive compensation and growth opportunities.

Citizen Development

In some cases, companies need to upskill their existing employees, such as by embracing the new trend of Citizen Development—a new paradigm of no-code development.

Citizen Development does not require programming skills but enables individuals to design, develop, and deploy applications into production environments, much like experienced programmers. The trend of Citizen Development is driven by no-code platforms, which provide visual interfaces and drag-and-drop tools, allowing users without programming backgrounds to create applications.

No-code lowers the barrier to entry, enabling broader participation in digital transformation. Meanwhile, AI has numerous applications in software development, including code generation, testing, and debugging. By integrating no-code with AI, businesses and individuals can build applications more easily. Additionally, AI can enhance the capabilities of no-code platforms, helping to identify and fix potential issues and vulnerabilities.

The application scenarios for AI and no-code are extensive, spanning customer service, finance, healthcare, logistics, and more.

For example, no-code platforms can be used to develop AI-powered chatbots that provide instant customer service to website visitors. No-code platforms can also create financial reports or healthcare applications, helping doctors and patients efficiently manage health conditions and streamline patient care processes.

Undeniably, AI holds immense potential but also has its limitations. While these limitations are not insurmountable, they highlight the need for continued education, research, and development to ensure that AI can be effectively applied across various industries. This will usher in a new era of Citizen Development.

Author: Noel Carroll

Noel Carroll is an Associate Professor of Business Information Systems at the University of Galway and the founder of the Citizen Development Lab.