Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Companies expect to bear more expenses as a result of poorly implemented autonomous systems.
SHAPCHARGE | E+ | Ghetto images
Artificial intelligence opportunities are evolving rapidly and companies worldwide are fiercely trying to maintain and apply AI tools, but there are consequences for negligent performance.
In fact, 79% of companies worldwide expect to take on “AI debt” as a result of poorly applied autonomous instruments, according to a new report of Asana of AI’s condition at work who examined over 9,000 knowledge workers in the US, UK, Australia, Germany and Japan.
The report emphasizes that companies are unprepared and have no infrastructure and supervision needed to promote smooth cooperation between human employees and autonomous AI agentsS Other than generative AI, agents act independently, can initiate actions and recall the previous work they have done. Some examples include Openai’s operator and the Claude of Anthropic.
The debt of AI is the price for the proper implementation of nascent autonomous systems, said Mark Hoffman, an expert at the ASANA innovation laboratory, CNBC Make It told CNBC.
“These costs can be money for money. They can also be lost time, which is about money. This can also be a lot of things to cancel, which is expensive from a financial point of view. It burns people to do it. These are all the costs related to poor application,” Hoffman said.
The report outlines that debt can be manifested as security risks, poor data quality, low -impact agents that will waste time and resources for human employees, and a difference in management skills.
Hoffman said this is not an exhaustive list and the “debt” may look like a bunch of code created by AI that does not work properly or generated by AI content that no one uses.
New research from the Betterup laboratories and the Social Media Laboratory in Stanford even found that 40% of US Bureau workers have received AI-GENERATED “WORKSLOP” “ which researchers have described as a content that looks good but lacking any substance.
It has been created almost two hours of extra work for people who have encountered it, $ 186 invisible tax per month and $ 9 million per year, according to the study.
“There is currently big investments in this space and it is ultimately a question whether these investments will be paid,” Hoffman said.
Henry Aider, founder of the AI consulting firm, LaTent Space Advisory, and an adviser to the United Kingdom Government, Meta and Ai Video Startup Synthesia, emphasized the need for deliberate execution and structures.
“People who are CTOS or innovative employees, the good ones I have worked with, those I think I have made the best position to succeed with it, they are not a sugar coating of the interruption, that it will cost … As with any major processing, you will have problems, you will have bumps on the road,” Aider said in an interview.
“This is not a magic silver bullet”
Asana’s report found that despite the adoption of AI, increasing to 70% in 2025, from 52% in 2024, workers were also facing higher levels of digital burning.
Digital exhaustion increased to 84% in 2025 from 75% in the previous year, while unmanageable loads also increased to 77%, according to the report.
Mona Mourshed, based on a Global CEO of Generation -based employment organization, told CNBC that although companies are deploying AI instruments and promoted IT, workers are still struggling.
“The main reason that they are fighting and we know this from conversations with our own graduates is that the case of using how and why you should use this AI tool in the flow of your work is often missing,” Murshed said.
“Without a clear understanding of what the case is, which will make this specific task better, faster, more cheap … This leads to exhaustion because you do not know what the result is,” she added.
Mourshed noted that companies invest in AI in the hope that work at night will be better, faster and more cheaper, but they do not offer the necessary training or guidance to enable improvements.
“This is not a magic silver bullet and suddenly does everything you want after installing it … It will be a much more painful journey to get to these benefits than the companies that thought it.”
The Ai Ajder expert said the correct strategy carefully tests the use of AI and the construction of AI and builds an infrastructure around it, not to rush into the race unprepared.
“You do not start just with the installation, you start with piloting, you start from the range, from sandboxes, trying these systems,” he said.
This includes everything from proper training for employees to thinking about the type of AI models that business may need. It is much more difficult to answer mistakes or malfunctions when there is no procedure.
“So I’m not saying that you can’t take the risk thoughtfully when it comes to using AI, but it must be calculated and it must be covered,” Ajder said.
