Generative AI in Software Development: Boosting Productivity While Raising Concerns

Gen AI could speed up coding, but businesses should still consider risks

Developers can now turn to generative artificial intelligence (Gen AI) to code faster and more efficiently, but they should do so with caution and no less attention than before. While the use of AI in software development may not be new -- it's been around since at least 2019 and Gen AI brings significant improvements in the generation of natural language, images, and -- more recently -- videos and other assets, including code, Diego Lo Giudice, Forrester's vice president, and principal analyst, told ZDNET. Previous iterations of AI were used mostly in code testing, with machine learning leveraged to optimize models for testing strategies, Lo Giudice told ZDNET. Gen AI goes beyond these use cases, offering access to an expert peer programmer or specialist (such as a tester or a business analyst) who can be queried interactively to find information quickly. Gen AI can also suggest solutions and test cases. "For the first time, we are seeing significant productivity gains that traditional AI and other technologies have not provided us with," he said. Also: Can AI and automation properly manage the growing threats to the cybersecurity landscape? By 2026, the IDC predicts that 40% of new applications in Asia-Pacific will be "intelligent apps," which integrate Gen AI to enhance user experience and deliver new use cases. Specifically, the top impact of using Gen AI in software development and design will be increased productivity and developer speed, according to the IDC study, commissioned by the low-code development platform, OutSystems. Such potential business benefits will drive 30% of Asia-Pacific organizations to pay a premium of between 11% and 20% for Gen AI features in app development, the IDC study projects. The research firm further anticipates that, by 2025, more than 60% of businesses worldwide will adopt low-code technologies to facilitate faster application delivery and enhance operational efficiency. Also: Why data is the Achilles Heel of AI (and every other business plan) Developers can tap AI across the entire software development lifecycle, with a dedicated "TuringBot" at each stage to enhance tech stacks and platforms, Lo Giudice noted. Forrester coined TuringBots to describe AI-powered tools that help developers build, test, and deploy code. The research firm believes TuringBots will drive a new generation of software development, assisting at every stage of the development lifecycle, including looking up technical documentation and auto-completing code. "Analyze/plan TuringBots," for instance, can facilitate the analysis and planning phase of software development, Lo Giudice detailed, pointing to OpenAI's ChatGPT and Atlassian Intelligence as examples of such AI products. Others, such as Google Cloud's Gemini Advanced, can generate designs of microservices and APIs with their code implementation, while Microsoft Sketch2Code can generate working code from hand-written sketched UI, he said. Also: The data suggests gen AI boosts software productivity - for these developers Lo Giudice added that "coder TuringBots" are currently the most popular use case for Gen AI in software development, generating code from prompts as well as from code context and comments via autocompletion for popular integrated development environments (IDEs). These include common languages such as JavaScript, C++, Python, and Rust. A big draw of generative models is that they can write code in many languages, allowing developers to input a prompt to generate, refactor, or debug lines of code, Michael Bachman, Boomi's head of architecture and AI strategy, said. "Essentially all humans interacting with Gen AI are quasi and senior developers," he said. The software vendor integrates Gen AI into some of its products, including Boomi AI, which translates natural language requests into action. Developers can use Boomi AI to design integration processes, APIs, and data models to connect applications, data, and processes, according to Boomi. The company uses Gen AI to support its own software developers, who closely watch the code that runs its platform. Also: Gen AI as a software quality tool? Skepticism is fading as more organizations implement it "And that is the key," Bachman said. "If you are using Gen AI as the primary source for building your whole application, you are probably going to be disappointed. Good developers use Gen AI as a jumping-off point or to test failure scenarios thoroughly, before putting code into production. This is how we deal with that internally." His team also works to build features to meet their customers' "practical AI objectives." For example, Boomi is creating a retrieval system because many of its clients want to replace keyword searches with the ability to look up content, such as catalogs on their websites, in natural language. Developers can also use Gen AI to remediate security, Lo Giudice said, looking for vulnerabilities in AI-generated code and offering suggestions to help developers fix certain vulnerabilities. Also: Can AI and automation properly manage the growing threats to the cybersecurity landscape? Compared to traditional coding, a no- or low-code development strategy can offer speed, built-in quality, and adaptability, Forrester principal analyst John Bratincevic said. It also provides for an integrated software development lifecycle toolchain and access to an expanded talent pool that includes non-coders and "citizen developers" outside the IT community, Bratincevic said. However, organizations may face challenges related to the governance of large-scale implementation, especially with managing citizen developers who can number in the thousands, he cautioned. Pricing can also pose a barrier, as it is typically based on the number of end users, he said. While Gen AI or AI-infused software assistants can enable junior professionals to fill talent gaps, including in cybersecurity, Lo Giudice said an expert eye review is still necessary for all these tasks. Also: AI accelerates software development to breakneck speeds, but measuring that is tricky Bratincevic concurred, stressing the need for developers and other employees in the software development lifecycle to review everything the platform generates or auto-configures using AI. "We are not yet, and probably won't ever be, at the point of trusting AI blindly for software development," he said. For one, there are security requirements to consider, according to Scott Shaw, Thoughtworks' Asia-Pacific CTO. The tech consultancy regularly tests new tools to improve its efficiency, whether in the IDE or to support developers work. The company does so where it is appropriate for its customers and only with their consent, Shaw told ZDNET, noting that some businesses are still nervous about using Gen AI. "Our experience is that [Gen AI-powered] software coding tools aren't as security-aware and [attuned with] security coding practices," he said. For instance, developers who work for organizations in a regulated or data-sensitive environment may have to adhere to additional security practices and controls as part of their software delivery processes. Also: Businesses still ready to invest in Gen AI, with risk management a top priority Using a coding assistant can double productivity, but developers need to ask if they can adequately test the code and fulfill the quality requirements along the pipeline, he noted. It's a double-edged sword: Organizations must look at how Gen AI can augment their coding practices so the products they develop are more secure and, at the same time, how AI brings added security risks with new attack vectors and vulnerabilities. Because it delivers significant scale, Gen AI amplifies everything an organization does, including the associated risks, Shaw noted. A lot more code can be generated with it, which also means the number of potential risks increases exponentially. While low-code platforms may be a good foundation for Gen AI Turingbots to aid software development, Bratincevic noted that organizations should know which large language models (LLMs) are used and ensure they align with their corporate policies. He said Gen AI players "vary wildly" in this respect, and urged businesses to check the version and licensing agreement if they use public LLMs such as OpenAI's ChatGPT. Also: How to use ChatGPT to write code: What it does well and what it doesn't He added that Gen AI-powered features for generating code or component configurations from natural language have yet to mature. They may see increased adoption among citizen developers but are unlikely to impress professional developers. Bratincevic said: "At the moment, a proven and well-integrated low-code platform plus Gen AI is a more sensible approach than an unproven or lightweight platform that talks a good game on AI." While LLMs carry out the heavy lifting of code writing, humans still need to know what is required and provide the relevant context, expertise, and debugging to ensure the output is accurate, Bachman added. Developers also need to be mindful of sharing proprietary data and intellectual property (IP), particularly with open-source tools, he stated. They should avoid using private IP, such as code and financial figures, to ensure they are not training their Gen AI models using another organization's IP, or vice versa. "And if you choose to use an open-source LLM, make sure it is well-tested before putting it into production," he added. Also: GitHub releases an AI-powered tool aiming for a 'radically new way of building software' "I would err on the side of being extremely circumspect about the models that Gen AI tools are trained on. If you want those models to be valuable, you have to set up proper pipelines. If you do not do that, Gen AI could cause a lot more problems," he cautioned. It is early days and the technology continues to evolve; its impact on how roles -- including software developers -- will change remains far from certain. For example, AI-powered coding assistants may change how skills are valued. Shaw quipped: will developers be deemed better because they are more experienced or because they can remember all the coding sequences? For now, he believes the biggest potential use case is Gen AI's ability to summarize information, offering a good knowledge base for developers to better understand the business. Then, they can translate that knowledge into specific instructions, so systems can execute the tasks and build the products and features customers want.

Gen AI gives software developers surge in productivity - but it's not for everyone

Generative AI brings productivity benefits, but IT professionals must be wary of wholesale adoption of these tools in their current incarnations. For anyone building software, generative AI (Gen AI) -- especially a tool like GitHub Copilot -- is a means to quickly create, test, document, and debug code, which leads to big productivity benefits. This boost frees up the time, resources, and brainpower of software developers and operations professionals to step up and fill consultative and leadership roles within their organizations. Also: 4 ways to turn generative AI experiments into real business value However, while the productivity benefits are clear, AI may not benefit everyone, and industry experts advise treading cautiously into automation. First, some context. Gen AI code-suggestion tools can boost software developer productivity, according to a multi-party study by researchers at Microsoft, MIT, Princeton University, and the University of Pennsylvania. The research analyzed the output of 4,867 software developers across three companies, all with access to Copilot, and discovered a 26% productivity increase in the weekly number of completed tasks, along with a 14% increase in the number of code updates, and a 38% increase in the number of times code was compiled. Other executives and professionals across the industry agreed Gen AI can produce a massive productivity boost for software developers. "Gen AI and copilot tools are significantly impacting development velocity," Brett Smith, a distinguished software developer with SAS, told ZDNET. "AI can help write boilerplate code, unit tests, and documentation, freeing the developer to accelerate solving the actual solutions. Generative AI has unquestionably revolutionized the game for software development, serving as a pair programmer to developers worldwide." Also: Technologist Bruce Schneier on security, society and why we need 'public AI' models Generative tools can be useful for various development tasks, such as "introducing new features to existing codebases or porting the codebases to new programming languages," said Flavio Villanustre, global chief information security officer at LexisNexis Risk Solutions. "These AI tools also help with the software archeology that is necessary to understand poorly documented codebases. Additionally, AI can serve as a helpful resource when test cases must be created -- especially in test-driven software development environments. Last, but not least, AI tools can help identify third-party libraries and frameworks that could be useful in particular projects." However, it's not all good news -- and the researchers in the Microsoft/MIT/Princeton/UPenn study highlighted one major caveat: the benefits of Gen AI diminish among developers with greater experience. "Less-experienced developers showed higher adoption rates and greater productivity gains," they stated. "Copilot significantly raises task completion for more recent hires and those in more junior positions, but not for developers with longer tenure and in more senior positions." Also: Perplexity AI's new tool makes researching the stock market 'delightful'. Here's how AI-driven tools "are incredibly useful for less-experienced developers," agreed Edward White, head of growth at beehiiv, a digital newsletter service. "They offer real-time suggestions for refactoring and optimization, guiding junior developers through best practices while they code. These tools can identify inefficiencies or repetitive patterns and recommend improvements, making the code cleaner and more efficient. The instant feedback helps developers learn how to write maintainable code, follow proper naming conventions, and use better structures." However, despite the focus on less-experienced developers, Smith said long-time professionals can also see the benefits of Gen AI. "In my experience, veteran developers have greatly benefitted from AI assistance," he said. "AI is incredibly efficient at writing boilerplate code, and it frees the developer to do the complex bespoke things that AI is not good at. In general, developers with less experience struggle with solving complex problems, and AI is usually unable to help them in that aspect." Also: The secret to successful digital initiatives is pretty simple, according to Gartner The evidence from industry experts suggests a balance must be achieved. Whether Gen AI is deployed by experienced or inexperienced developers, IT professionals and executives must be wary of wholesale adoption of these tools in their current incarnations. "While the assistant is typically able to write certain common functions, interface reasonably with existing libraries, write test cases, and explain existing code, it does make certain mistakes that a proficient programmer would avoid," said Villanustre. "These AI-based tools also have challenges when addressing more complex algorithms and can write code that is unsafe or insecure. In my opinion, current AI tools are commensurate with an entry-level programmer, and still require monitoring by more experienced software professionals." Smith also said the increased velocity from Gen AI tools is a double-edged sword. "Developers write bad code; AI helps them write bad code faster," he said. "The increase in the amount of code a team produces could introduce more bugs and vulnerabilities than normal. This wave of new code can quickly overwhelm testing and security teams. Teams will need to automate checkpoints, testing, and security scanning to keep pace in fighting the evil aspects of generative AI moving forward." Also: CIOs must also serve as chief AI officers, according to Salesforce survey The quality of the generated code is also an issue. "AI is only as good as its training," said David Brault, an expert at Mendix. The training data may include "a combination of well-written and substandard code." This mix might lead to code of varying quality and consistency and can even build on technical debt. Another challenge with AI-generated code is the risk of incompatibility with existing systems, especially complex or legacy architectures. "While AI tools can produce efficient code for specific tasks, they may not always consider the unique dependencies, frameworks, or structures of older systems," White cautioned. "This mismatch can lead to problems such as unexpected behavior or even cause disruptions if the AI-generated code is implemented without thorough testing." Also: Generative AI doesn't have to be a power hog after all To reduce these issues, "developers should carefully evaluate the compatibility of AI-generated code with their current infrastructure," White advised. "This includes conducting extensive testing in controlled staging environments and, if needed, customizing the code to fit the particularities of their system." IT professionals must also be wary of legal issues that can crop up along the way. "Using AI-generated code requires caution regarding duplication, intellectual property, and potential licensing issues," said Mira Nathalea, chief marketing officer of SoftwareHow, a provider of software reviews. "Developers and companies should be mindful of the risks associated with unverified code suggestions. There's the risk of unknowingly incorporating code that may not align with the project's licensing requirements, possibly resulting in intellectual property conflicts. Developers should review AI-generated suggestions carefully to maintain code quality and legal compliance." Also: Gen AI as a software quality tool? Skepticism is fading as more organizations implement it Software professionals and executives must also carefully manage the ethical concerns that can arise from AI-based code generation. "Overlooking ethical considerations can lead to unintended issues, such as privacy risks, data misuse, or biases within algorithms," said Joel Popoff, CEO of Axwell Wallet, a maker of tech-friendly everyday carry items. "When AI-generated code isn't carefully evaluated, it may not fully comply with regulations or ethical standards, especially in areas like user privacy and equitable treatment." Popoff said it's important to implement ethical review procedures tailored to AI-generated code to mitigate such risks. "This includes regularly auditing AI outputs, ensuring that data is handled responsibly, and actively checking for bias in automated decisions," he said. "Establishing transparency about how AI processes and uses data helps maintain trust and keeps the development process aligned with ethical principles." Also: 3 ways to build strong data foundations for AI implementation, according to business leaders There are, therefore, a range of issues and concerns that must be considered. However, when employed responsibly, AI might revolutionize the development profession. "One of the biggest advantages of using AI in development is the ability to automate the tedious and repetitive tasks that developers hate as they waste time and add little business value to the solution," said Brault. "Instead, by automating these low-level routines with AI, developers can spend more time finding creative ways to solve the complex business problems that impact their company. If anything, developers will become even more valuable as their expertise will be required to validate and secure AI-generated code as well as build reusable components, templates, and frameworks to establish governance standards for integrating AI with their existing systems."

AI-driven software testing gains more champions but worries persist

The question is no longer whether gen AI and emerging technologies will elevate the automation of software testing but when and how. However, proponents need to overcome organizational resistance. Comprehensive quality engineering and testing are a must for today's software-driven organizations. Perhaps not surprisingly, generative artificial intelligence (Gen AI) is emerging as a cutting-edge component of the quality and testing phase of the software development lifecycle. However, long-term success in software-testing automation is about establishing the necessary organizational will and resources. In short, to paraphrase management guru Peter Drucker's oft-cited phrase: Culture eats software-quality strategies for breakfast. Also: The data suggests gen AI boosts software productivity - for these developers "The debate on which quality engineering and testing activities will benefit most from Gen AI remains unresolved," said the co-authors of an OpenText study involving 1,755 tech executives state. The survey, released by Capgemini and Sogeti (part of the Capgemini Group), pointed to a growing focus on leveraging Gen AI "for test reporting and data generation over test-case creation." AI creates an answer, or at least a partial answer, to many nagging software quality issues. Software quality has been a challenge since the first computers were built eight decades ago, and in a world awash in technology networks and solutions, the problem has only grown more acute. Gen AI is emerging as an important step in managing quality. Also: 6 ways to write better ChatGPT prompts - and get the results you want faster The survey confirmed about seven in ten organizations (68%) employ Gen AI to assist with their software quality efforts. At least 29% of organizations have fully integrated Gen AI into their test automation processes, while 42% are exploring its potential. The study also suggested that "cloud-native technologies and robotic process automation, with Gen AI and predictive AI both playing significant roles" are prevalent in this new area of test automation. "Cloud-native technologies are appealing because they open the door to cost-effective solutions that eliminate the need for tooling licenses, which lowers overall operational expenses. It is no longer a question of 'if' AI and other emerging technologies will become a part of the DevOps fabric. We are in the early stages of a dynamic shift in the way we do business." Also: Google's new AI course will teach you to write more effective prompts - in 5 steps The conclusion is that AI represents the next stage of automation for relatively complex quality assurance and testing processes. "There is a clear need to align quality engineering metrics with business outcomes and showcase the strategic value of quality initiatives to drive meaningful change," the survey's team of authors, led by Jeff Spevacek of OpenText, stated. "On the technology front, the adoption of newer, smarter test automation tools has driven the average level of test automation to 44%. However, the most transformative trend this year is the rapid adoption of AI, particularly Gen AI, which is set to make a huge impact." Spevacek and his co-authors continued: "The evolution of large language models and AI tools, particularly Copilot, have enabled their seamless integration into existing software development lifecycles, ushering in a new wave of efficiency and innovation in quality engineering automation." In the previous year's software quality survey, "we saw an uptick in the investments made by organizations in AI solutions to drive the quality-transformation agenda," they wrote. "However, a significant number were skeptical about the value of AI in quality engineering." Also: Think AI can solve all your business problems? Apple's new study shows otherwise Attitudes toward AI have shifted significantly over the past 12 months: "A large number of organizations are now moving [away] from experimenting to real-scale implementation of Gen AI to support quality engineering activities. We truly believe we will see further advancements in this area." However, employing AI as a software quality assurance tool is challenging. At least 61% of survey respondents said they worry about data breaches associated with leveraging generative AI solutions. A lack of comprehensive test automation strategies and a reliance on legacy systems were identified by 57% and 64% of respondents, respectively, as key barriers to advancing automation efforts. The picture is also mixed for embedding quality engineers with Agile software delivery teams. Only one-third of respondents said most of their quality engineers participate in Agile teams. However, the authors suggested this lack of participation might not be a bad thing. "This suggests a growing recognition of the need for quality engineers who can operate independently of Agile teams, while still contributing to overall quality objectives. In fact, the number of standalone quality engineers is expected to increase from 27% to 38%." Also: Google survey says more than 75% of developers rely on AI. But there's a catch The survey suggested this increase in high-quality engineers may also reflect a trend of cross-skilling of Agile teams to address software quality and testing: "The focus on cross-skilling to align quality engineers more closely with Agile teams appears to have paid off. This year's survey results show that organizations have made considerable progress in upskilling their teams -- only 16% of respondents now view a lack of skills as a major bottleneck, a significant improvement from last year's 37%." However, despite this progress, most tech executives said there isn't enough emphasis on quality engineering. More than half (56%) said the challenge is that "quality engineering is not seen as a strategic activity in our organization." A similar proportion of respondents agreed that the "quality engineering process is not automated enough," and that "quality engineers lack the skillset to support Agile projects." Also: Agile development can unlock the power of generative AI - here's how The rise of Gen AI and predictive AI may offer a cost-effective and streamlined approach to aligning quality and testing efforts with overall software development and deployment. Some of the recommendations offered by the OpenText/Sogeti team for moving forward with automation and AI in software quality efforts included the following: While AI offers great promise as a quality and testing tool, the study said there are "significant challenges in validating protocols, AI models, and the complexity of validation of all integrations. Currently, many organizations are struggling to implement comprehensive test strategies that ensure optimized coverage of critical areas. However, looking ahead, there is a strong expectation that AI will play a pivotal role in addressing these challenges and enhancing the effectiveness of testing activities in this domain." The key takeaway point from the research is that software quality engineering is rapidly evolving: "Once defined as testing human-written software, it has now evolved with AI-generated code." Also: Could AI make data science obsolete? As a result of this evolution, quality engineering is seeing an increased volume of code and test scripts that need to be generated, and there are new requirements for testing software chains from end to end.