By Ellie Stewart
Job seekers looking to increase their AI skillset in order to land their next role may be building AI skills that hiring professionals aren’t necessarily looking for.
New research by Adobe Acrobat reveals that there is a gap between the skills related to AI that people are acquiring and those that employers really value. Moreover, the idea that people should get more technical knowledge about AI appears to overlook the value of soft skills like judgment, teamwork, and critical thinking.
According to the Adobe data, the most popular technical skill among job seekers was brainstorming using AI, which accounted for 30% of surveyed job seekers. Then comes AI image generation and workflow automation, 19% each. But only 45% of people reported feeling comfortable with their skills to write prompts.
When it comes to recruiters, understanding where this gap exists can be important for hiring the best talent.
As for ethical AI supervision, it was chosen by 33% of hiring managers, whereas creative intuition was prioritized by 28%. Especially notable is that hiring managers prioritized ethical AI supervision 74% more often than prompt engineering.
This difference may acquire more significance as the availability of generative AI tools decreases its value.
In case almost everyone can ask the AI system to generate the picture, summarize a document or come up with some ideas, building a skillset that not only utilizes AI tools but thoughtfully partners with them may be what helps a job seeker stand out from the crowd.
According to the Adobe data, the biggest red flag related to AI in an interview process was a candidate's inability to check facts, cited by 70% of hiring professionals
Also, nearly half thought that taking full credit for work done with AI help is something to be worried about.
The first generation of workplace AI skills was based on usage and execution. Could an employee use a chatbot? Could he create an image? Did she understand prompting?
These execution based skills now serve as a foundation for a larger skillset.
The most ideal candidates would show their understanding of how to analyze outputs, fix mistakes and pick tools, as well as build processes for using AI. This means that being able to explain the reasoning behind the use of an AI tool and what to do next is just as important as using the tool itself.
This may also change the tools employees choose to use. While many believe the best Ai tools to be those that generate the most output, there are also tools that can provide meaningful assistance throughout other steps of a project’s process. These tools may allow users to better optimize their work and bridge the gap between their skills and the tools’ capabilities.
This illustrates a much more teamwork-based approach to AI tool use, where the user leans on technology as a creative partner but allows for increased responsibility for the final product.
According to Adobe, over 90% of hiring managers stated that their organization offered training in the use of AI skills; however, knowledge sharing remains inconsistent. Even when looking at technology companies, only one out of three companies has a formal system in place to share knowledge of AI.
Moreover, the training is primarily focused on efficiency. Workflow automation was the most prevalent AI skill taught, accounting for 52% of courses offered, followed by AI brainstorming (48%) and prompt engineering (32%). AI output auditing, even though employers pay so much attention to this aspect, accounts for up only 30%.
This leaves businesses with their own challenge.
Organizations, if they want employees to use AI responsibly, question its outputs and have a good judgment regarding AI outputs, should train employees on such matters rather than offer them access to more tools.
According to the Adobe survey, nearly three in five hiring professionals state that AI mastery improves an employee’s job security. However, over 60% also agreed that they may still hire a candidate without basic AI proficiency.
This means that AI skill building in the workplace may be changing directions. The optimal AI user at work is someone who is capable of using AI technologies in a way that compliments their human skills, especially when it comes to decision making.
For job seekers, this means highlighting your ability to implement AI tools and explaining the thought process behind your work. For organizations working towards future workplace solutions, it means utilizing AI tools where they best support your workers and enhance their own skillsets.
The survey included 805 job seekers and 406 hiring professionals, with fieldwork conducted in late April 2026.
Reviewed by Irfan Ahmad.
Read next: Online conflict may strengthen, not weaken, social movements, McGill researchers find
Job seekers looking to increase their AI skillset in order to land their next role may be building AI skills that hiring professionals aren’t necessarily looking for.
New research by Adobe Acrobat reveals that there is a gap between the skills related to AI that people are acquiring and those that employers really value. Moreover, the idea that people should get more technical knowledge about AI appears to overlook the value of soft skills like judgment, teamwork, and critical thinking.
According to the Adobe data, the most popular technical skill among job seekers was brainstorming using AI, which accounted for 30% of surveyed job seekers. Then comes AI image generation and workflow automation, 19% each. But only 45% of people reported feeling comfortable with their skills to write prompts.
When it comes to recruiters, understanding where this gap exists can be important for hiring the best talent.
The new AI skillset expands beyond the prompt
Hiring managers prioritized time management above all other soft skills, with 47% selecting it, followed by adaptive problem solving at 44% and collaboration at 41%.As for ethical AI supervision, it was chosen by 33% of hiring managers, whereas creative intuition was prioritized by 28%. Especially notable is that hiring managers prioritized ethical AI supervision 74% more often than prompt engineering.
This difference may acquire more significance as the availability of generative AI tools decreases its value.
In case almost everyone can ask the AI system to generate the picture, summarize a document or come up with some ideas, building a skillset that not only utilizes AI tools but thoughtfully partners with them may be what helps a job seeker stand out from the crowd.
According to the Adobe data, the biggest red flag related to AI in an interview process was a candidate's inability to check facts, cited by 70% of hiring professionals
Also, nearly half thought that taking full credit for work done with AI help is something to be worried about.
AI fluency is becoming a process, not a software skill
The current working landscape has changed what it means to be skilled with AI tools.The first generation of workplace AI skills was based on usage and execution. Could an employee use a chatbot? Could he create an image? Did she understand prompting?
These execution based skills now serve as a foundation for a larger skillset.
The most ideal candidates would show their understanding of how to analyze outputs, fix mistakes and pick tools, as well as build processes for using AI. This means that being able to explain the reasoning behind the use of an AI tool and what to do next is just as important as using the tool itself.
This may also change the tools employees choose to use. While many believe the best Ai tools to be those that generate the most output, there are also tools that can provide meaningful assistance throughout other steps of a project’s process. These tools may allow users to better optimize their work and bridge the gap between their skills and the tools’ capabilities.
This illustrates a much more teamwork-based approach to AI tool use, where the user leans on technology as a creative partner but allows for increased responsibility for the final product.
Employers have a skills gap of their own
It is not the sole responsibility of employees to bridge the gap.According to Adobe, over 90% of hiring managers stated that their organization offered training in the use of AI skills; however, knowledge sharing remains inconsistent. Even when looking at technology companies, only one out of three companies has a formal system in place to share knowledge of AI.
Moreover, the training is primarily focused on efficiency. Workflow automation was the most prevalent AI skill taught, accounting for 52% of courses offered, followed by AI brainstorming (48%) and prompt engineering (32%). AI output auditing, even though employers pay so much attention to this aspect, accounts for up only 30%.
This leaves businesses with their own challenge.
Organizations, if they want employees to use AI responsibly, question its outputs and have a good judgment regarding AI outputs, should train employees on such matters rather than offer them access to more tools.
The most valuable AI skill might be knowing what humans still need to do
The suggestion that technical understanding of AI is irrelevant does not hold much water.According to the Adobe survey, nearly three in five hiring professionals state that AI mastery improves an employee’s job security. However, over 60% also agreed that they may still hire a candidate without basic AI proficiency.
This means that AI skill building in the workplace may be changing directions. The optimal AI user at work is someone who is capable of using AI technologies in a way that compliments their human skills, especially when it comes to decision making.
For job seekers, this means highlighting your ability to implement AI tools and explaining the thought process behind your work. For organizations working towards future workplace solutions, it means utilizing AI tools where they best support your workers and enhance their own skillsets.
The survey included 805 job seekers and 406 hiring professionals, with fieldwork conducted in late April 2026.
Reviewed by Irfan Ahmad.
Read next: Online conflict may strengthen, not weaken, social movements, McGill researchers find
