You're sitting in your quarterly team meeting when your manager opens a slide deck and announces something big: the company is rolling out a new AI platform. It’s designed to help automate repetitive tasks, generate faster insights, and lighten your workload. One tool automatically pulls data from internal systems and formats it into weekly reports. Another suggests customer responses based on past conversations. It all sounds promising, at least in the moment.

A few demos follow. There’s some polite enthusiasm, maybe a raised eyebrow or two. Then the meeting ends, and the platform disappears into the background.

Two weeks later, you’re back to manually copy-pasting numbers into Excel. 

This kind of rollout is more common than it seems. Across industries, companies are investing in AI with the hope of transforming productivity. However, without a clear plan for implementation—or meaningful investment in training—the technology often stalls before it even starts. Untethered from the workflows they were meant to improve, AI tools risk becoming optional extras rather than practical upgrades.

Even when teams want to make use of AI, the environment can work against adoption. People might not know what the tools are supposed to replace. Access to the right data may also be missing altogether. In some organizations, information is scattered across separate platforms, locked behind strict permissions, or stored in formats that don’t work across regions or AI systems. Instead of forming a shared knowledge base, the data becomes fragmented, duplicated, and difficult for AI tools to analyze and synthesize. 

Oftentimes, what looks like resistance by employees is really something else: hesitation grounded in ambiguity. Employees may not ignore AI’s potential because they’re unwilling to learn, but because they’re uncertain about its relevance, application, or how to begin. Without that clarity, even the most advanced platform might feel like one more tool to work around—not with.

AI can still reshape how we work, but adoption takes more than access. It takes intention, coordination, and trust that these tools were made to support and not replace the work people already do.

Artificial Intelligence (AI): The ability of software to carry out tasks that traditionally required human intelligence, mirroring cognitive functions usually associated with human minds. 

Deep Learning: A subset of machine learning that relies on deep neural networks, which are layers of connected “neurons” whose links have parameters or weights that can be trained. Deep learning powers many AI applications today, including image recognition, voice assistants, and predictive analytics.

Large language models (LLMs): A foundational model class that can process massive volumes of unstructured text and learn relationships between words or sub-word tokens. This allows LLMs to produce natural-sounding language and complete tasks like summarizing or extracting information. GPT-4, which underlies ChatGPT, is one example of an LLM.

Symbolic AI: An early form of artificial intelligence based on manually coded rules and logic structures that simulate human reasoning. These systems relied on symbolic representations to process information but lacked the ability to learn from new data or experience.

Data Fragmentation: A challenge that arises when data is spread across multiple systems and locations, such as cloud platforms or on-premises servers, preventing organizations from fully leveraging AI. For example, an employee well-being report might be compiled differently across departments, making it difficult for AI tools to generate a consistent, organization-wide view.

Long before artificial intelligence was a field, it was fiction. In the late 1800s, writers began imagining machines that could think, speak, and sometimes rebel.1 These early stories weren’t based on real technology. They reflected a growing curiosity about whether machines could one day think like people.

One of the most influential early computing machines was the Electronic Numerical Integrator and Computer (ENIAC), built in 1945 at the University of Pennsylvania by engineers John Mauchly and J. Presper Eckert.² Designed to calculate artillery trajectories for the U.S. Army, the machine could complete in seconds what once took humans hours. It didn’t think, but it computed data fast, reliably, and on a scale that had never been seen. For many scientists, it was the first time they could picture a future where machines could follow instructions and possibly make decisions.

In 1950, British mathematician Alan Turing posed a now-famous question: “Can machines think?” His paper, “Computing Machinery and Intelligence”, offered a practical test.³ Instead of debating definitions, Turing proposed an experiment, known as the Turing Test, which places a human and a computer behind separate screens and lets a judge converse with both.⁴ If the judge couldn’t reliably tell them apart, could we say the machine was intelligent? The Turing Test became an enduring thought experiment, shaping decades of research and sparking debate about the nature of cognition, language, and understanding.

Five years later, a group of researchers had gathered at a major conference to further explore what Turing had implied about the possibility of engineering intelligence. At that conference, mathematician John McCarthy coined the term “artificial intelligence” and later developed LISP, a programming language designed to support what’s called symbolic AI.⁵ These systems, however, couldn’t learn from experience or datasets to answer novel questions. 

Symbolic systems began making their way into real-world settings during the early 1980s, with one standout system, called XCON, that helped configure computer hardware for customers and automated decisions that once took hours of human labor.⁶ Still, these systems had limits. They couldn’t adapt. Rules had to be written by hand, and unexpected inputs often caused them to fail. Parallel to research in expert systems, another approach was being studied. Inspired by the brain’s architecture, early neural networks tried to simulate how neurons fire, connect, and learn from repetition.⁷ These systems weren’t based on rules. They relied on adjusting internal weights—numerical values controlling the strength of connections between artificial neurons—as they processed data.

In the early 2000s, there were several breakthroughs in AI research. More powerful processors, better training algorithms, and unprecedented volumes of data gave neural networks a new life. Among the researchers leading this charge was Geoffrey Hinton, whose research team introduced a model called AlexNet at a global image recognition competition in 2012.⁸ AlexNet could scan millions of photos and trained itself to identify patterns that no one had explicitly taught it. The outcome redefined AI’s core methods, moving the field firmly into the era of deep learning, which refers to the use of multilayered neural networks that learn complex patterns from vast datasets through repeated exposure and adjustment.

From there, momentum in implementing AI within public settings had surged. Neural networks powered advances in speech recognition, language translation, and recommendation engines. In 2015, OpenAI was founded to steer this rapidly evolving field toward socially beneficial outcomes.⁹ Five years later, the organization released GPT-3, a language model with 175 billion parameters.¹⁰ It could write essays, summarize reports, generate code, and respond to prompts in plain English. The outputs weren’t flawless, but they were coherent, fast, and very humanlike.

By 2022, AI systems powered by neural networks had rapidly integrated into everyday infrastructure. In healthcare, they flagged irregular X-ray scans and assisted with early diagnoses.¹¹ In the manufacturing industry, they predicted equipment failures before downtime occurred.¹² In customer service, they fueled chatbots that could handle complex questions with minimal handoffs.¹³

Still, widespread adoption often collides with organizational barriers. In many workplaces, AI tools remain underutilized—not because they fall short, but because the path to integration is murky.^14,15 Skills gaps can slow momentum, especially when teams lack the training to use these systems confidently. Cultural resistance is just as significant. For some, AI sounds like a threat, something that might edge out their role rather than lighten the load.¹⁶ For others, it disrupts a routine they’ve spent years perfecting. 

Trust compounds the challenge. These systems process immense volumes of personal and professional information that need to be protected from privacy violations or misuse. In no other field is the ethical compass more relevant. Computer scientist Fei-Fei Li is a prominent figure known for her emphasis on integrating ethical principles into the design of AI to benefit humanity and mitigate potential harms.¹⁷ Adopting AI in the workplace means rethinking not just software, but structure: how do we move past these barriers to ensure it serves the greater good?

John Mauchly and J. Presper Eckert

American physicist John Mauchly and electrical engineer J. Presper Eckert collaborated to invent the Electronic Numerical Integrator and Computer (ENIAC) in the 1940s, which is widely regarded as the first general-purpose electronic digital computer.² Together, this duo of scientists helped lay the technical and conceptual groundwork that has since shaped the field of computer science and AI. 

Alan Turing 

British mathematician and computer scientist who is best known for developing the Turing Test.⁴ He conducted some of the earliest work in the field of AI during the mid-20th century and had proposed one of the first formal frameworks to assess whether a machine could demonstrate intelligence.

John McCarthy 

A computer scientist and mathematician, McCarthy developed LISP in the 1950s, which is a programming language that models human reasoning through symbolic manipulation. He’s also regarded as one of the founders of the discipline of AI and coined the term in 1956.⁵

Geoffrey Hinton

A British cognitive psychologist and computer scientist who revolutionized the field of AI by implementing neural networks into its study.⁷ Throughout his career, Hinton has been dedicated to building robust computational models of perception and memory, a commitment that earned him the 2024 Nobel Prize in Physics for his foundational contributions to artificial neural networks.

Fei-Fei Li

Chinese-American computer scientist whose work today aims to broaden the scope of AI research by championing ethics-centered design.¹⁷ She’s been a major force in shaping the responsible use of AI into a central concern, pushing the field to weigh social impact as seriously as technical innovation.

AI is no longer a futuristic concept: it’s a present-day tool with very real implications. Although the technology itself has made rapid advancements, many organizations are still stalled at the adoption stage. That disconnect can carry serious consequences, not just for innovation, but for care, efficiency, and competitiveness.

AI hesitation puts companies at risk

AI might be on everybody’s lips, but in many organizations, it’s still stuck in the planning stage. Some teams are eager but lack technical fluency. Others have the tools, but no clear framework for how to use them.18 What’s more, concerns about privacy, security, or job displacement often slow momentum before it even begins.¹⁹ 

That hesitation comes at a cost.

According to Thomson Reuters’ 2025 Future of Professionals report, companies with clear AI strategies—ones that move beyond trial and error—were over three times more likely to report tangible benefits, from faster service delivery to measurable revenue growth.²⁰ Despite this, only 22% of businesses reported having such a plan.

That gap matters. In a separate study, more than 700 highly skilled consultants were asked to complete a complex creative task: pitch a new product idea for a shoe company, map out its development and launch, and write a detailed report analyzing the entire process.²¹ Participants were split into three different groups. Some were given access to GPT-4, others were not. Among those who used the tool, performance increased by 38%. When users were also shown a brief guide on how to engage with GPT-4 effectively, the improvement climbed to 42.5%.

The takeaway here isn’t just that AI can enhance human work—it’s that the upside depends on how well systems are integrated and supported. Without training, without guidelines, and without leadership to steer adoption, even the most powerful tools might remain underused. In industries that move quickly, that kind of inertia doesn’t just slow progress, but it can jeopardize a company’s relevance altogether. When competitors are harnessing AI to boost creativity, streamline operations, and deliver results faster, staying still becomes the bigger risk.

How hospitals can benefit from further AI integration

Many hospitals are already contending with rising patient volumes, mounting documentation demands, and ongoing staffing shortages.²² These challenges make the integration of AI appear rather daunting. However, these same pressures are exactly what make incorporating AI within healthcare settings worth the effort. When deployed effectively, AI can help hospitals lighten their load—reducing administrative drag, easing the burden of data entry, and guiding clinical decisions in ways that feel faster, steadier, and in some cases, more precise than current routines allow.²³

Emma‑Jane Spencer, a researcher in AI medicine, highlights a challenging paradox in one critical healthcare department, the intensive care unit (ICU).²⁴ Although ICUs generate massive volumes of real-time physiological data, AI remains largely absent from how that information is actually used. This bottleneck, however, isn’t technological; it’s organizational.

A key culprit for why AI is underutilized in healthcare settings is data fragmentation. Patient information lives in disconnected systems, scattered across platforms that don’t speak to one another.²⁴ As a result, even when hospitals license a predictive model, it might not have the inputs it needs to function. On top of that, legal and privacy concerns frequently stall the implementation of AI tools. Hospital data are some of the most sensitive out there, and without robust governance structures in place, many institutions would rather sideline innovation than risk reputational or compliance fallout.

These organizational barriers can come with a major price. For example, sepsis—a life-threatening condition in which the body’s immune response spirals out of control—requires rapid detection and immediate treatment. Recent models show that AI can flag the warning signs of sepsis within the first hour of ICU admission.²⁵ By analyzing subtle changes in ECGs, blood pressure, and other vital sign waveforms, these systems surface risks long before lab results arrive. That head start could mean quicker antibiotics, more precise decisions, and a better shot at recovery.

Plenty of hospitals already have the data. Some even have the tools. What’s missing is a framework that turns capability into care—one that catches what matters right when it matters most. Without that connection, promising models stay on the sidelines, while the pressures on frontline staff keep piling up.

AI hesitation may stall the hiring process 

A surge in job applications might seem like a good problem to have—unless your organization lacks the right tools, or trust in them. 

According to a recent survey conducted by Canadian recruitment firm Robert Walters, over 70% of hiring managers have noticed a substantial rise in job applicants, while more than half also report that this increase has been bogging down their hiring process.²⁶ Simultaneously, nearly 40% of job seekers are submitting more than 20 applications a week, flooding inboxes and straining HR teams already stretched thin.

In this kind of volume, even the most promising resumes can get lost in the shuffle. Without support, recruiters face an impossible task—discerning top talent while racing against deadlines, stakeholder pressure, and operational costs. It’s no surprise that hiring delays have a ripple effect: product launches stall, teams stay understaffed, and momentum dries up.

To keep things moving, some firms are investing in AI-powered screening tools. In one study by Ujlayan et al. (2023), a system trained on over 1,000 tech-sector job seekers used profile mapping and similarity analysis to align candidates with job descriptions.²⁷ The result? Manual screening time fell by 80%. That’s the kind of intervention that makes hiring a much more efficient process.

Still, uptake remains cautious. In a survey of 238 recruiters, more than 70% voiced concerns that AI lacks the judgment, empathy, and contextual awareness humans bring.²⁸ Some feared that over-relying on algorithms trained on biased or outdated data could cause more harm than good.

These hesitations are more than philosophical; they’re organizational barriers. When tools like AI are available but mistrusted or poorly implemented, businesses can lose ground. Talent won’t wait. In a market where people are your greatest asset, overlooking a great candidate is a cost few companies can afford.

You might assume that introducing AI into the workplace is a matter of technical readiness. Often, however, the deeper challenge is how it disrupts trust. From biased hiring tools to opaque diagnostic systems and fears of job loss, the integration of AI can reinforce existing inequalities and heighten uncertainty across industries.

When AI in hiring reinforces outdated biases

Although AI is often pitched as a solution to human error in hiring, that narrative may start to crack when the technology itself introduces new kinds of bias. For organizations eager to modernize their recruitment processes, these risks aren’t just theoretical. They’re structural barriers that slow adoption, stall trust, and raise ethical red flags from the outset.

Kyra Wilson, a doctoral student at the University of Washington’s Information School, recently led a study probing whether AI models used in hiring might disadvantage candidates based on race and gender.²⁹ She and her team modified more than 550 resumes and 570 job descriptions, then tested them against three open-source large language models (LLMs) from Salesforce, Contextual AI, and Mistral, tools commonly embedded in HR pilots to parse job descriptions, match resumes, and rank candidates.

The results were as uncomfortable as they were illuminating. Even when qualifications such as education and experience were held constant, candidates with White-associated names were favored 85% of the time.²⁹ Resumes with female-associated names received top placement only 11% of the time. In some cases, these models appeared to favor White men for roles that have never historically skewed male or White,  suggesting that the bias wasn’t simply reflecting old hiring data. It was shaping new, algorithm-driven patterns of discrimination that hadn’t previously existed. 

This presents a serious hurdle for businesses considering AI-powered hiring systems. Until AI tools can demonstrate fairness across intersecting identities in organizational settings, skepticism won’t just linger, but also stall adoption and raise the stakes for every organization that chooses to implement these systems despite these risks.

AI-assisted medical diagnoses might be ready, but are clinicians?

Hospitals have long been at the forefront of innovation. With AI now capable of flagging anomalies in scans, streamlining patient intake, and even identifying rare diseases, the potential for impact in healthcare is enormous.^23,24 Yet the real test of AI adoption doesn’t lie in what the software can do. It lies in who’s willing to use it.

While many clinical settings have embraced AI in administrative roles, its integration into decision-making remains tentative. A 2019 study by Cai and colleagues offers a telling glimpse into why.³⁰ When 21 pathologists were introduced to a deep learning model for diagnosing prostate cancer, they didn’t just ask whether the model got the answer right. They wanted to know how it worked, what kinds of cases confused it, and whether it had blind spots. Most of all, they wanted clarity on its underlying objective—what the system was designed to prioritize in the first place.

In a more recent national survey, 147 Turkish oncologists echoed these concerns.³¹ Although nearly 80% had used AI tools like LLMs in some capacity, fewer than 10% reported receiving formal training. The mismatch between exposure and expertise was striking. Even with broad enthusiasm for using AI to support prognosis and research, many respondents hesitated when it came to ethical risks, patient communication, or the possibility of weakened trust in the physician’s role. 

The oncologists also presented possible solutions. Over 80% had advocated for national or international standards, and more than half supported the creation of new laws and dedicated oversight institutions. ³¹ Others called for informed consent clauses specific to AI use—clear disclosures embedded into patient forms, not buried in fine print.

Healthcare providers aren’t resisting change, but are calling for clarity. Without clear standards, formal training, and thoughtful regulatory policies, even the most promising tools may stall before reaching the bedside. For AI to truly support clinical care, adoption must be built on trust as much as on technical capabilities.

Fear of replacement stalls workforce productivity

One of the most widely cited fears about artificial intelligence—losing your job to it—is often dismissed as exaggerated. After all, most AI systems are meant to streamline tasks, not eliminate workers. Or so the story goes.

Findings from recent labor market research may suggest something more complicated. In a large-scale U.S. labor force analysis, Broady and colleagues examined which occupations were most exposed to automation, including AI-driven tools.³² The results pointed to a worrying trend: the more automatable a job was, the more at risk it was of being significantly altered. Such jobs were also more likely to be lower paid, physically demanding, and held by Black and Hispanic workers. Additionally, they often lacked basic safeguards: predictable hours, skill-building opportunities, and benefits that support long-term growth.³²

By contrast, positions less vulnerable to automation tended to offer better compensation, greater flexibility, and stronger buffers against economic shocks. Broady and colleagues note that, without intervention, existing race and income gaps may grow wider as automation expands.

There’s also the issue of perception. In a three-part study on Chinese office workers, Wu and colleagues found that fear of AI-related job loss had serious psychological consequences.³³ Their findings were telling: individuals who believed their roles were at risk reported higher anxiety and lower motivation to learn new tools. Rather than sparking innovation, the looming threat of replacement seemed to paralyze.

This issue raises a deeper concern. Although AI is often introduced as a tool to improve efficiency or streamline decisions, its presence in the workplace can carry unintended consequences. When workers feel dispensable, they may be less likely to take initiative, ask questions, or invest in new skills.

Educators navigate AI without clear direction

When conversations about AI in education surface, the spotlight often lands on students. Think plagiarism concerns, AI-generated essays, or questions about learning integrity. However, there’s another story unfolding—one that centers not on learners, but on those responsible for teaching them.

A growing number of educators are exploring ways to incorporate AI into their daily routines, from streamlining lesson plans to analyzing student progress.³⁴ Yet for many, adopting these tools still feels like navigating without a map.

A recent nationwide study conducted in the United States by the RAND Corporation, a major American nonprofit global policy research institute, uncovered some of the structural reasons why.³⁴ Drawing from the spring 2024 American Educator Panel surveys, the study revealed that while most U.S. principals and teachers have encountered AI tools, few have received meaningful training. 

Among the top barriers cited by principals nationwide were a lack of professional development (reported by 72% of respondents), concerns about data privacy (70%), and uncertainty about how AI fits into their professional responsibilities (also 70%).³⁴ Teachers echoed similar concerns. These weren’t one-off complaints but had revealed a deeper tension between technological opportunity and organizational readiness.

The divide was especially stark across socioeconomic lines. In high-poverty schools, principals were more likely to report receiving no district-level guidance on how to use AI in classrooms. Concerns about data use, quality of outputs, and the ethics of using AI for grading or content creation loomed large. For some, questions of plagiarism and professional integrity remained unresolved.

Interestingly, the U.S. Department of Education released a national AI integration toolkit in early 2024, encouraging schools to adopt evidence-informed strategies and prioritize AI literacy. Yet RAND’s findings suggest that guidance alone falls short. Without the infrastructure to deliver that guidance equitably, many educators remain unsure whether they’re crossing ethical lines. While interest in AI is high, many educators simply don’t feel equipped to use it with confidence.

The AI gap between employees and executives

Imagine being told that nearly every person at your workplace is using a new tool, but leadership hasn’t quite caught up. That’s not a hypothetical scenario. It’s what the multinational consulting firm McKinsey & Company found when surveying over 3,600 employees and 238 executives across the United States and five other countries.¹⁸ Nearly every company had invested in AI, yet fewer than 1% believed they were operating at full maturity. The real roadblock wasn’t employee resistance; it was leadership hesitation.

According to the data, C-suite leaders estimated that just 4% of employees use generative AI for at least 30% of their daily work. Employees, however, reported doing so at three times that rate. Additionally, while only 20% of leaders thought employees would use AI extensively within the next year, nearly half of employees believed they would.

In the study, McKinsey introduced four workplace archetypes: Zoomers, Bloomers, Gloomers, and Doomers—to capture how employees feel about AI. Surprisingly, even the most skeptical groups, like Gloomers and Doomers, reported high levels of comfort using generative AI at work. Gloomers, who advocate for stricter top-down regulation, and even the pessimistic Doomers weren’t avoiding AI altogether. In fact, over 90% of Gloomers and 70% of Doomers had already experimented with AI tools.¹⁸

What barriers stand in the way of effective AI use, then? A lack of clear direction. Nearly half of U.S. employees said they wanted formal training. Around 45% wished AI tools could be integrated more smoothly into their workflows, and over 40% simply wanted easier access to the tools. Yet one in five reported little to no visible support from leadership when it came to building AI-related skills.

International comparisons revealed a sharper contrast to U.S. employees. McKinsey’s data showed similar interest in using AI across employees in Australia, India, Singapore, New Zealand, and the UK. The difference? Infrastructure. While just over half of U.S. employees said their companies offered strong support for AI learning, that figure rose to 84% among their international peers. These workers weren’t only being trained—they were actively involved in shaping the tools through feedback, beta testing, and development requests.¹⁸ 

When leaders hesitate about the use of AI in the workplace, momentum stalls. McKinsey calls the missing link “superagency,” a state where employees are empowered, and leaders move decisively alongside them.¹⁸ However, without leadership stepping in to guide, support, and scale those efforts, AI’s potential in the workplace will keep hovering just out of reach.

Is Automation Coming for Your Job?

Worried that automation or AI will replace your role? You're not alone in that concern. Still, history and data indicate that’s unlikely. This article breaks down why entire jobs rarely disappear, and instead, how AI reshapes the tasks within them. 

Neural Networks

Neural networks are a subset of machine learning inspired by the structure of the human brain. By mimicking how neurons interact and learn from repeated experiences, neural networks can identify patterns, make predictions, and solve complex problems. This article explores how neural networks work, what they’ve made possible, and the challenges that come with scaling them responsibly.

1. Butler, S. (1863). Darwin among the machines. June, 13, 205.
2. Strauss, A. (1976). ENIAC: The first computer. Mathematics Teacher, 69(1), 66–72.
3. Turing, A. M. (1950). Computing machinery and intelligence. Mind, 59(236), 433–460. https://doi.org/10.1093/mind/LIX.236.433
4. Pinar Saygin, A., Cicekli, I., & Akman, V. (2000). Turing test: 50 years later. Minds and Machines, 10(4), 463–518.
5. Moor, J. (2006). The Dartmouth College Artificial Intelligence Conference: The next fifty years. AI Magazine, 27(4), 87–89.
6. Rashid, A. B., & Kausik, M. A. K. (2024). AI revolutionizing industries worldwide: A comprehensive overview of its diverse applications. Hybrid Advances, 7, 100277.
7. Hinton, G. E. (1992). How neural networks learn from experience. Scientific American, 267(3), 144–151.
8. Alom, M. Z., Taha, T. M., Yakopcic, C., Westberg, S., Sidike, P., Nasrin, M. S., ... & Asari, V. K. (2018). The history began from AlexNet: A comprehensive survey on deep learning approaches. arXiv preprint arXiv:1803.01164.
9. OpenAI. (2015, December 11). Introducing OpenAI. OpenAI. https://openai.com/index/introducing-openai/
10. Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J. D., Dhariwal, P., ... & Amodei, D. (2020). Language models are few-shot learners. Advances in Neural Information Processing Systems, 33, 1877–1901.
11. Hunter, B., Hindocha, S., & Lee, R. W. (2022). The role of artificial intelligence in early cancer diagnosis. Cancers, 14(6), 1524.
12. Fernandes, M., Corchado, J. M., & Marreiros, G. (2022). Machine learning techniques applied to mechanical fault diagnosis and fault prognosis in the context of real industrial manufacturing use-cases: A systematic literature review. Applied Intelligence, 52(12), 14246–14280.
13. Nirala, K. K., Singh, N. K., & Purani, V. S. (2022). A survey on providing customer and public administration based services using AI: Chatbot. Multimedia Tools and Applications, 81(16), 22215–22246.
14. Pagallo, U., O’Sullivan, S., Nevejans, N., Holzinger, A., Friebe, M., Jeanquartier, F., ... & Miernik, A. (2024). The underuse of AI in the health sector: Opportunity costs, success stories, risks and recommendations. Health and Technology, 14(1), 1–14.
15. Harish, V., Sharma, R., Rana, G., & Nayyar, A. (2023). Artificial intelligence in sustainable education: Benefits, applications, framework, and potential barriers. In The role of sustainability and artificial intelligence in education improvement (pp. 219–236). Chapman and Hall/CRC.
16. Chen, S., & Zhao, Y. (2025). Explore the AI anxiety attitude of E-commerce employees. International Journal of Human–Computer Interaction, 41(13), 8438–8446.
17. Bommasani, R., Arora, S., Chayes, J., Choi, Y., Cuéllar, M. F., Fei-Fei, L., ... & Song, D. (2025). Advancing science-and evidence-based AI policy. Science, 389(6759), 459–461.
18. Mayer, H., Yee, L., Chui, M., & Roberts, R. (2025, January 28). Superagency in the workplace: Empowering people to unlock AI’s full potential. McKinsey & Company. https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work
19. Moradi, P., & Levy, K. (2020). Displacement or risk-shifting? In The Oxford Handbook of Ethics of AI (p. 271). Oxford University Press.
20. Thomson Reuters. (2025, June 26). The AI adoption reality check: Firms with AI strategies are twice as likely to see AI-driven revenue growth; those without risk falling behind. https://www.thomsonreuters.com/en/press-releases/2025/june/the-ai-adoption-reality-check-firms-with-ai-strategies-are-twice-as-likely-to-see-ai-driven-revenue-growth-those-without-risk-falling-behind
21. Dell’Acqua, F., McFowland III, E., Mollick, E. R., Lifshitz-Assaf, H., Kellogg, K., Rajendran, S., ... & Lakhani, K. R. (2023). Navigating the jagged technological frontier: Field experimental evidence of the effects of AI on knowledge worker productivity and quality. Harvard Business School Technology & Operations Mgt. Unit Working Paper, (24-013).
22. Murphy, G. T., Sampalli, T., Bearskin, L. B., Cashen, N., Cummings, G., Rose, A. E., ... & Villeneuve, M. (2022). Investing in Canada’s nursing workforce post-pandemic: A call to action. Facets.
23. Durrah, O., Aldhmour, F. M., El-Maghraby, L., & Chakir, A. (2024). Artificial intelligence applications in healthcare. In Engineering Applications of Artificial Intelligence (pp. 175–192). Springer Nature Switzerland.
24. Spencer, E. J., Economou-Zavlanos, N. J., & van Genderen, M. E. (2025). What if we do, but what if we don’t? The opportunity cost of artificial intelligence hesitancy in the intensive care unit. Intensive Care Medicine, 51(2), 378–381.
25. Mollura, M., Lehman, L. W. H., Mark, R. G., & Barbieri, R. (2021). A novel artificial intelligence based intensive care unit monitoring system: Using physiological waveforms to identify sepsis. Philosophical Transactions of the Royal Society A, 379(2212), 20200252.
26. Robert Walters. (2024, September 25). Job application overload. Robert Walters Canada. https://www.robert-walters.ca/insights/news/blog/job-application-overload.html
27. Ujlayan, A., Bhattacharya, S., & Sonakshi. (2023). A machine learning-based AI framework to optimize the recruitment screening process. International Journal of Global Business Competitiveness, 18, 38–53.
28. Horodyski, P. (2023). Recruiter's perception of artificial intelligence (AI)-based tools in recruitment. Computers in Human Behavior Reports, 10, 100298.
29. Wilson, K., & Caliskan, A. (2024, October). Gender, race, and intersectional bias in resume screening via language model retrieval. In Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society (Vol. 7, pp. 1578–1590).
30. Cai, C. J., Winter, S., Steiner, D., Wilcox, L., & Terry, M. (2019). "Hello AI": Uncovering the onboarding needs of medical practitioners for human-AI collaborative decision-making. Proceedings of the ACM on Human-Computer Interaction, 3(CSCW), Article 104. https://doi.org/10.1145/3359206
31. Erdat, E. C., & Çay Şenler, F. (2025). Turkish medical oncologists’ perspectives on integrating artificial intelligence: Knowledge, attitudes, and ethical considerations. BMC Medical Ethics, 26(1), 82.
32. Broady, K. E., Booth-Bell, D., Barr, A., & Meeks, A. (2025). Automation, artificial intelligence, and job displacement in the US, 2019–22. Labor History, 1–17.
33. Wu, T. J., Liang, Y., & Wang, Y. (2024). The buffering role of workplace mindfulness: How job insecurity of human–artificial intelligence collaboration impacts employees’ work–life-related outcomes. Journal of Business and Psychology, 39(6), 1395–1411.
34. Kaufman, J. H., Woo, A., Eagan, J., Lee, S., & Kassan, E. B. (2025, February 11). Uneven adoption of artificial intelligence tools among U.S. teachers and principals in the 2023–2024 school year (RR-A134-25). RAND Corporation. https://www.rand.org/pubs/research_reports/RRA134-25.html