The AI industry has made major promises about its tech boosting the productivity of developers, allowing them to generate copious amounts of code with simple text prompts.<\/p>\n
But those claims appear to be massively overblown, as The Register reports<\/a>, with researchers finding that productivity gains are modest at best \u2014 and at worst, that AI can actually slow down human developers.<\/p>\n In a new report<\/a>, management consultants Bain & Company found that despite being \u201cone of the first areas to deploy generative AI,\u201d the \u201csavings have been unremarkable\u201d in programming.<\/p>\n \u201cGenerative AI arrived on the scene with sky-high expectations, and many companies rushed into pilot projects,\u201d the report reads. \u201cYet the results haven\u2019t lived up to the hype.\u201d<\/p>\n First off, \u201cdeveloper adoption is low\u201d even among the companies that rolled out AI tools, the management consultancy found.<\/p>\n Worse yet, while some assistants saw \u201cten to 15 percent productivity boosts,\u201d the savings most of the time \u201cdon\u2019t translate into positive returns.\u201d<\/p>\n It\u2019s yet another warning shot, highlighting concerns that even in one of the most promising areas, the AI industry is struggling to live up to its enormous hype. That\u2019s despite companies pouring untold billions of dollars into its development, with analysts openly fretting about an enormous AI bubble<\/a> getting closer to popping<\/a>.<\/p>\n In July, a damning study<\/a> by nonprofit Model Evaluation & Threat Research found that AI coding tools may not just realize the expected productivity gains, they could slow software developers down<\/a>.<\/p>\n \u201cSurprisingly, we find that when developers use AI tools, they take 19 percent longer than without \u2014 AI makes them slower,\u201d the researchers wrote. That\u2019s because hallucinations forced the developers to spend extra time cleaning up and reviewing code. The tools also performed worse when dealing with large and complex code repositories, often misunderstanding the context.<\/p>\n Programming website Stack Overflow also found in a survey<\/a> earlier this year that developers\u2019 trust in AI tools had cratered \u2014 despite a rise in the percentage of developers using them. Many of them pointed to solutions that are \u201calmost right, but not quite.\u201d<\/p>\n \u201cOne of the most surprising findings was a significant shift in developer preferences for AI compared to previous years, while most developers use AI, they like it less and trust it less this year,\u201d Stack Overflow senior analyst for market research and insights Erin Yepis told VentureBeat<\/a> in July.<\/p>\n \u201cThis response is surprising because with all of the investment in and focus on AI in tech news, I would expect that the trust would grow as the technology gets better,\u201d she added.<\/p>\n It remains to be seen whether the advent of \u201cagentic AIs,\u201d designed to autonomously execute a series of tasks, will change the situation.<\/p>\n \u201cAgentic AI is already reshaping the enterprise, and only those that move decisively \u2014 redesigning their architecture, teams, and ways of working \u2014 will unlock its full value,\u201d the report reads.<\/p>\n Besides slowing coders down, experts have warned that AI coding assistants could introduce major security problems<\/a> as well. In a recent report<\/a>, security firm Apiiro found that developers who use AI produce ten times more security problems than their counterparts who don\u2019t use the tech.<\/p>\n