The integration of AI at work raises concerns about its impact on cognition and collaboration. Historical anxieties about technology eroding mental capacity, from Socrates' critique of writing to fears of digital dependency, reappear in current debates. While AI promises efficiency, it reshapes how individuals think and interact. Recent work on "System 0" describes a AI-mediated layer of cognition preceding Kahneman's dual-process model, complicating notions of ownership, expertise, and responsibility. This intersects with distributed cognition (Hutchins, 1995), activity theory (Engeström, 2001), and extended mind theory (Clark & Chalmers, 1998), which view cognition as distributed across people and artifacts. AI intensifies this extension, but also introduces "competence penalty", professionals judged less capable when their work is linked to AI, even if quality remains unchanged (Gai, 2025). We conducted a review of AI-related cognitive outsourcing at work and its effects on workplace dynamics. Cognitive outsourcing is conceptually broader the "cognitive offloading" and is foundational to distributed cognition and transactive memory systems. Cognitive automation is a newer concept driven by advances in AI/ML, related to knowledge work and human-AI collaboration. Our review shows that cognitive outsourcing impacts the nature of work, affecting job design, skill requirements, and organizational performance, reshaping work by automating tasks, shifting skill requirements, and altering job meaning. A parallel phenomenon is "workslop," low-effort and low-quality AI output lacking contextual understanding (Niederhoffer, 2025). In collaborative environments, this creates a "cognitive commons" problem: individuals maximize efficiency but collective outcomes decline. At the same time, competence penalty pressures workers to obscure AI use, producing tensions between reliance and credibility. These dynamics weaken trust, mentorship, and peer learning, as AI-mediated content lacks the subtle markers of human reasoning that sustain professional growth. Future directions should focus on developing synergies between human and AI cognition, exploring hybridization processes that sustain expertise while amplifying organizational innovation.