New report says AI has only added another layer of work for employees
Advocates of artificial intelligence tools promised workplaces that this would free employees' time to pursue more high-value work, but a new report says that they might have forgotten to mention the manual trade-off that is offsetting the time saved from AI.
Research from Connext Global's 2026 AI Oversight Report revealed that there is still a "hidden aftermath layer" when using AI tools in the workplace.
According to the report, 42% of AI users said they still need to edit or fix the output produced by AI before sending or using it at work.
This comes as only 37% of employees reported that AI is right without fixes most of the time, while 63% said it is only right sometimes or less.
"That gap has direct implications for productivity," the report reads.
Nearly half (46%) of employees using AI said fixing takes about the same time as doing the work manually, while 11% said it takes more time.
"In other words, 57% report that once correction is required, the time advantage can disappear, reshaping ROI for everyday tasks," the report read.
AI use in HR
The findings come in the wake of widespread use of AI in organisations, including the HR department.
Common applications of generative AI in HR include drafting job descriptions and role summaries, creating guides and structured question sets, and summarising interview notes or application materials, according to talent data platform Findem. Other applications in HR include:
- Building templates for nurture campaigns in CRMs
- Generating reminders, rejections, and thank-you messages for candidates
- Producing training materials and internal communications
- Preparing high-level summaries of hiring activity or process improvement ideas
But even the output produced by AI tools still needs finishing touches from humans amid widespread risks, according to previous research.
Output produced by AI tools is still first drafts that need human expertise and require reviewing and editing to ensure it reflects the company's voice and values.
AI also risks adding outdated language or unintentionally biased phrasing, requiring HR leaders to review the output to ensure these do not occur.
Prevent additional workload from AI
Connext Global noted that its findings do not mean AI does not deliver speed.
"AI can accelerate output, but most teams are still responsible for making it accurate, complete, and ready for real-world use," said Tim Mobley, president and CEO of Connext Global.
According to the report, the best outcomes appear when teams match AI to tasks where review is straightforward, the acceptable margin of error is clear, and the workflow is designed so humans can quickly validate AI output.
"Without that structure, the fix cycle can become its own workload," the report read. "When fixes are routine, the quality bar moves from generation to verification and that verification becomes part of the job."
Mobley stressed that the solution is to build these "oversight habits" which can maintain the high quality of AI output.
"The opportunity is not just adopting AI, it is building the oversight habits that keep quality high while speed increases," he said.
