AI Slop: The 'Plausible Trash' Spreading Through Organizations

Yoon Myung-hoon, Head of Operations at WantedLab, defined AI slop as 'plausible but content-less reports created by employees with little desire to do well, using AI.' These reports demand significant time and effort from leaders to review 'plausible but pretty trash' and verify their accuracy. In this process, leaders must pay a hand-off cost for the review.

Head of Operations Yoon Myung-hoon classified AI-based decision-making types into four categories based on the presence or absence of questions and answers. First, cases where both the question and answer are known, such as understanding sales status, which involves clear data. Second, cases where a question exists but the answer is unknown, such as market forecasting or strategic analysis, where AI's predictive capabilities are crucial. Third, cases where an answer exists but no question, which involves utilizing tacit knowledge within the organization that is not clearly formalized. Finally, cases where neither the question nor the answer is known, representing a stage of exploring and discovering entirely new domains.

Tacit knowledge refers to personal know-how or experiences that vaguely exist within an organization but are not clearly codified. Head of Operations Yoon Myung-hoon pointed out that this tacit knowledge is the area where AI adoption yields the least effect. To transform tacit knowledge into an organizational asset, the process of converting meeting minutes into text and recording them for AI agent learning is essential. Leaders must focus on acquiring know-how generated during work processes through communication with team members and converting it into explicit knowledge.

With a dramatic increase in employees' work processing speed after AI adoption, leaders are facing different management styles than in the past. Head of Operations Yoon Myung-hoon analogized that if it felt like working with 'ten people strolling' in the past, now it feels like working with 'ten racehorses.' This means that if a leader looks away for even a moment, the direction of work can deviate unexpectedly. Therefore, the leader's role must evolve beyond simple management to 'orchestration' – observing all channels and maximizing work synergy.

Head of Operations Yoon Myung-hoon emphasized that AI ROI (Return On Investment) should be understood not as financial return, but as the 'compound interest of intelligence' (Return Intelligence). He explained that if leaders only focus on short-term financial gains, they will struggle to experience the true effects of AI adoption. The value of AI manifests in continuously teaching tacit knowledge within the organization, accumulating intelligence through this process, and fostering long-term growth.

Evaluating AI investment performance solely based on short-term revenue can hinder the compound effect of intelligence. Given the future uncertainties AI may bring, it is important to utilize leading indicators that can identify what elements appear in terms of value matching, rather than focusing solely on complete outcome metrics. In other words, indicators that predict and measure how AI is creating new value within the organization and what positive changes are occurring in that process are crucial.

Recent research suggests that a problem exists where negligent employees continue to generate AI slop. Poorly made outputs using AI, or slop, incur significant hand-off costs for leaders during the review process. Especially when the reviewer is not the actual author of the report, it takes more time and effort to verify its authenticity and understand its content. This exacerbates unnecessary workload within the organization and is a major cause of overall inefficiency.

Head of Operations Yoon Myung-hoon stressed that leaders in the AI era must clearly present purposes and goals, while granting team members autonomy in 'how-to'—meaning the methods of task execution. This autonomy implies the authority for team members to find the optimal methodology and perform tasks independently. The leader's role has become crucial in setting the direction for achieving final goals, and trusting and supporting team members to reach those goals in creative and efficient ways.

Head of Operations Yoon Myung-hoon suggested a 'wicked problem'-centric approach to motivate employees. This method involves sharing challenging problems that no one has solved yet with team members and motivating them through the process of leaders and employees finding answers together. This prioritizes a collaborative relationship of deliberation and problem-solving over an authoritative knowledge transfer method from the leader. He explained that it is important to agree on goals with team members and participate in joint problem-solving, under the premise that even leaders can be wrong.

Head of Operations Yoon Myung-hoon argued that the core of performance management should be 'Discovery.' He distinguished between 'operational efficiency,' which solves existing known problems, and 'discovery' of new facts. He stated that a key performance indicator should be whether new 'discoveries' were made this quarter or month, meaning whether new material for the next strategy has been secured. He pointed out that performance management focusing on a single indicator can limit the organization's diverse possibilities, and penny-pinching reward policies can be counterproductive, emphasizing new attempts and creative exploration.