You don’t have to look far to find predictions about the industries and product categories that generative AI technologies seem to disrupt. One niche that is safe from ChatGPT and others, at least for now, is low-code software development platforms.
Low-code development enables both trained and citizen developers to efficiently create applications regardless of their background in coding and software development practices. There are many low-code platforms, each offering different capabilities for application design and development.
ChatGPT and similar AI technologies can generate code on demand. Due to ChatGPT’s lack of functionality and security, we don’t expect it to replace low-code development tools in most organizations anytime soon. In fact, AI enhances rather than replaces low-code platforms.
Similarities between ChatGPT and low-code
ChatGPT and AI-powered coding-specific technologies like GitHub Copilot and Amazon CodeWhisperer can handle some of the same tasks as low-code tools.
Automatically generate code tailored to user requests. For example, when ChatGPT was asked to create a Python function to retrieve data from her MySQL database, here is an excerpt of the response.
import pymysqldef pull_data_from_mysql(database, table, user, password,
host="localhost", port=3306):
"""
Connects to a MySQL database and pulls data from a specific table.
Returns the data as a list of tuples.
"""
# Connect to the database
conn = pymysql.connect(
host=host,
port=port,
user=user,
password=password,
database=database
)# Create a cursor
cur = conn.cursor()# Execute a SELECT statement
cur.execute(f"SELECT * FROM {table}")# Fetch all the rows
rows = cur.fetchall()# Close the cursor and the connection
cur.close()
conn.close()# Return the data as a list of tuples
return rows
Once you’ve included the appropriate values in this auto-generated code, your actual Python application will connect to your MySQL database.
By automatically generating code like this example, AI tools such as ChatGPT can be used by both non-technical business users who cannot code and professional software developers who need to generate code quickly. may be able to assist you. Low-code vendors also serve these groups.
Where ChatGPT falls short compared to low-code
Tools like ChatGPT, at least in their current state, don’t work as full-blown replacements for low-code platforms. These platforms have too many features that AI services can’t reproduce well.
Features that ChatGPT struggles to deliver include application design, code integration, security protection, and license management.
Application design
ChatGPT and similar tools can generate code that implements specific parts of your application. They are not very useful for selecting functional elements to include in an application or organizing those functionalities into a consistent application architecture.
The low-code tool’s visual interface shows you what features you can implement and how they might be integrated to create your application. Unlike ChatGPT, low-code tools help users think about their applications in the big picture and explore different possibilities for application design.
code integration
A low-code platform combines application components to form a complete application. Users select the features they want to implement and the tool automatically integrates them.
In contrast, general-purpose generative AI platforms such as ChatGPT can only write code in snippets. The user must decide how to combine these snippets into a functional application. Most non-technical users have no expertise in code integration at all. Professional developers can perform that integration, but that defeats the purpose of using AI to accelerate application development.
Code security
The code contained in the low-code development module is designed to be secure. Low-code vendors build security features into their application runtimes to detect risks and potential exploit attempts.
In contrast, ChatGPT and other general-purpose AI systems do not guarantee the safety of the generated code. These tools are unaware of the wider codebase into which their code may be integrated. So, for example, it might not consider how user input collected by one function is passed to another function. They don’t realize how their input could potentially inject malicious commands into their application.
For these reasons, ChatGPT can generate code vulnerable to injection attacks, cross-site scripting, insecure dependencies, and various other issues. These security risks exist to some extent in tools designed to generate code, such as low-code platforms.
license permission
ChatGPT and similar technologies are partially trained on open source code. The developers of that code did not consent to the training process, raising allegations that these AI tools, and in some cases the code they generate, violate open source licenses. Claims like these have begun to lead to lawsuits against the likes of GitHub Copilot.
It’s too early to tell whether an organization using ChatGPT to help create software could get into legal trouble. This risk does not exist for licensed low-code development platform customers.
AI capabilities in low-code tools
For now, AI will complement existing workflows rather than replace low-code development. AI-powered code implementations and interfaces are still underutilized in low-code platforms.
Indeed, it is possible to imagine a low-code development platform similar to ChatGPT. For example, a low-code vendor can add a natural language interface to their tools so that users can describe the kind of application they want to build. Low-code tools leverage libraries of pre-built functions to implement applications.
These vendors may continue to offer the option of building low-code applications through a visual interface that shows available functionality and allows users to combine it. Adding this AI-based feature allows users to make requests using natural language, providing a more comprehensive, secure, and legally sound application component and design tool than the code generated by ChatGPT. will be
