• By quade
• 25 July 2024

AI Trust Declines Among Developers Despite Increased Usage

Recent surveys have shown that AI usage among developers and businesses has been increasing over the past few years. However, these surveys also show that AI trust is shaky and has even declined in some places.

The issues many people have with AI involve its accuracy, capabilities, and potential risks. Many concerns are so great that it reveals a major gap between adoption and trust across nearly 200 countries.

AI Trust Survey Results

The survey in question was the Developer Survey launched by Stack Overflow which they conduct every year. This time, they asked over 65,000 respondents in 185 countries. All respondents are developers with experience in the fields AI is meant to play a role.

According to Ryan Polk, Chief Product Officer at Stack Overflow the results are clear AI remains a powerful tool in many workplaces, people are starting to understand it is not the catch-all solution that tech companies have advertised it to be. They conducted these surveys, to understand the developers’ thoughts and what changes AI trust.

They do this in hopes of improving some of the issues surrounding AI while minimizing their ethical and practical issues.

According to the survey, 76% of developers admit to using AI tools for work or plan to do so soon. That is a huge increase of 70% from 2023. But while there is an increase in adoption, fewer people trust AI as their favorability went down from 72% to 77% last year. Meanwhile, AI trust continues to remain low at 43%.

However, even with this increase in adoption, 45% acknowledge that AI tools are unreliable for complex tasks.

Beyond practical problems with AI, many ethical issues also affect AI trust. Among the most cited issues are problems with AI creating misinformation, ingrained bias from limited viewpoints, and poor citations. 

Surprisingly, one issue that developers did not have with AI is a fear that they may threaten their jobs. A large majority, about 70% believe AI will not be able to replace them.

AI Trust and Favorability Across the World and Other Results

Based on the country breakdown, AI favorability is highest in India, Spain, Portugal, and France which has favorability in the low to mid-70s. Meanwhile, countries that show lower favorability are Ukraine, Germany, and the UK which have favorability in the lower 60s.

Aside from the AI tools themselves, another important aspect is the use of coding tools. Today, JavaScript remains the dominant tool, followed by HTML/CSS, Python, SQL, and TypeScript. This is important as many report to using online sources to learn code and one of AI’s new goals is to be an automated coder.

Because of that, AI favorability is higher for Integrated Development environments with the highest being Visual Studio Code, followed by ChatGPT, GitHub, and Google Gemini.

How Does This Survey Affect BPOs

For Global BPO service providers, these surveys give a clue to what their clients and partners want. For example, the results of AI trust and favorability show whether or not it is pragmatic to use AI tools and whether they should consider traditional methods to provide services.

Meanwhile, they also offer indications of which tools people are willing to trust. This shows that established tools like ChatGPT remain the top players in AI trust and favorability. Meanwhile, newer tools still need to improve their services.

Lastly, it shows the areas where AI needs to improve and how BPO services can capitalize on these problems. From combating misinformation to providing more diverse viewpoints on certain topics.