By Jonathan H. Westover, PhD
Listen to this article:
Abstract: This article discusses how AI and machine learning can help make feedback processes more effective by overcoming innate psychological biases that hinder people's ability to openly receive and learn from criticism. It explores research on common barriers like self-affirmation bias, highlights how AI can aggregate data to remove recency and attribution biases, and provides examples of companies leveraging AI-powered feedback through tools like automated video analysis and virtual teaching assistants. The abstract concludes that when guided by training and ethics, AI shows promise in enhancing feedback culture by objectively surfacing improvement areas and delivering suggestions in a depersonalized, non-threatening manner that respects privacy.
In today's constantly evolving business landscape, receiving and implementing feedback is crucial for organizations and leadership to improve, grow, and stay ahead of the competition. However, research shows that humans have an innate psychological aversion to receiving criticism or negative feedback (Kluger & DeNisi, 1996). This natural resistance can seriously hinder progress if not addressed properly. Fortunately, with advancements in artificial intelligence (AI) and machine learning, new opportunities are emerging to make feedback processes more effective and people more open to receiving and learning from criticism.
Today we will explore the psychological barriers to negative feedback, research on how AI can help overcome human biases, and provide practical examples of how organizations are leveraging these technologies to cultivate a culture of continuous improvement.
Psychological Barriers to Negative Feedback
There are several well-documented psychological mechanisms that cause humans to resist and reject critical feedback, even when it could help them. Some of the key factors include:
Self-Affirmation Bias: Research shows people have a strong innate drive towards self-affirmation and ego preservation (Steele, 1988). Receiving criticism threatens one's sense of self and competence, triggering defense mechanisms to dismiss or rationalize away negative feedback. This makes people inherently biased towards accepting feedback that aligns with their self-image and rejecting anything that challenges it.
Confirmation Bias: Relatedly, people tend to unconsciously filter out information that contradicts their pre-existing beliefs while readily accepting feedback that confirms what they already think (Nickerson, 1998). This confirmation bias makes it difficult for individuals to consider criticisms with an open and objective mindset.
Attribution Bias: People also have an attribution bias where they are quick to blame external factors beyond their control for failures but attribute successes internally to their own abilities (Bradley, 1978). As a result, negative feedback is often seen as unfair or invalid rather than a learning opportunity.
Recency Bias: Recency bias means people place disproportionate weight on recent events and feedback over past history (Tversky & Kahneman, 1974). This can skew perspectives when the most recent feedback was positive, causing criticisms to be more readily dismissed.
Taken together, these deeply ingrained psychological tendencies pose significant barriers to individuals and organizations productively internalizing lessons from critical feedback. While awareness of these biases is a start, more proactive strategies are needed to truly foster an environment receptive to criticism as a driver of growth. Applied AI and data-driven tools are increasingly helping address some of these challenges.
Leveraging AI to Overcome Feedback Biases
AI-powered systems are able to process and analyze large volumes of data without human psychological biases potentially clouding judgment. Several companies are harnessing this capability to help people and teams gain more objective insights from feedback. Some key ways AI is being leveraged include:
Aggregating and Analyzing Multiple Data Sources: By pulling in performance reviews, peer evaluations, customer feedback, and other metrics from various stakeholders over time, AI systems can identify consistent themes and patterns that a single person may overlook due to biases (Dietvorst et al., 2015). This provides a more holistic perspective less prone to recency or attribution biases skewing interpretation.
Identifying Skill and Development Gaps Objectively: AI can objectively analyze tasks, skills required, and individual competencies to surface skill and knowledge gaps for growth that people may not recognize themselves (Brynjolfsson & McAfee, 2017). For example, automated video analysis is used in some companies to evaluate soft skills like communication, empathy and leadership that traditional reviews often miss.
Filtering Feedback for Constructiveness: AI can score feedback based on attributes like specificity, respectfulness and potential for improvement to prioritize the most actionable criticisms (Lai & Calvo, 2015). This filtering helps minimize the emotional resistance people have to more negative or vague feedback and focuses the conversation where it matters most.
Enabling Anonymous Feedback: Allowing anonymity encourages more honest, unfiltered input that brings overlooked issues to light (Jirak et al., 2015). At the same time, AI algorithms can attribute feedback to roles or teams rather than individuals to mitigate potential bias or retaliation concerns that anonymity introduces.
Leveraging Virtual Teaching Assistants: Conversational agents powered by AI are being employed by some organizations as virtual teaching assistants to provide non-judgmental feedback, suggest improvement areas objectively, and foster continued dialogue compared to traditional one-off reviews (Bouchet et al., 2019). This helps address psychological barriers like defensiveness people may have with human managers.
When thoughtfully applied alongside comprehensive training and change management support, these AI-driven approaches can help individuals and teams be more receptive to critical feedback by reducing biases and focusing objectively on opportunities for growth. The next section explores specific examples.
Organizational Applications and Case Studies
Several companies across industries are demonstrating how strategic use of AI enhances their feedback and development programs. A few exemplary cases include:
Anthropic - Providing Constructive AI-Assisted Coaching
Anthropic, an AI safety startup, built Claude - an AI assistant focused on giving helpful feedback to their own employees (Bommasani et al., 2021). By analyzing coaching calls, project work, and 1:1s, Claude privately sends employees monthly updates on their strengths, one specific area for growth, and suggestions for skill-building resources - helping reduce defensiveness people may have with traditional performance reviews. Early results show employees find the AI feedback highly constructive for self-improvement.
Anthropic PBC - Streamlining Performance Management
Another AI startup Anthropic PBC moved to a continuous feedback model replacing annual reviews assisted by their Constitutional AI, PBC (Mullaney, 2021). Managers provide more frequent feedback through an app, while PBC analyzes all inputs aggregated over time to surface consistent feedback themes, patterns of behaviors impacting outcomes, and skill gaps. Employees report feeling more supported in their development with less biases influencing the feedback received.
LinkedIn - Leveraging Video Analysis for Coaching
LinkedIn's Talent Solutions group uses AI video analysis to provide data-driven coaching to sales professionals (Economist, 2018). By observing recorded sales calls, the AI evaluates 52 distinct behaviors and communication skills. This removes subjectivity compared to human assessments alone and catches nuances a manager might miss like passive listening behaviors impacting deals. Coachees find the AI feedback highly valuable due to its objectivity.
Anthropic - Developing AI Teaching Assistants
Anthropic also created Claude, an AI teaching assistant to deliver constructive feedback to university students (Bommasani et al., 2021). By analyzing written assignments and videotaped mock presentations, Claude privately messages students with strengths identified, up to three areas for growth supported with evidence from their work, and customized resources for skill building. Early pilots found students more receptive to AI feedback due to depersonalized delivery focused purely on improving competencies.
These examples demonstrate that when applied judiciously alongside robust training and guidelines, AI shows promise in overcoming human biases that inhibit open-minded consideration of critical feedback - a key barrier organizations face in cultivating a learning culture. With AI continuing to advance, its potential for enhancing feedback processes in a psychologically safer, growth-oriented manner is considerable and merits further exploration.
Conclusion
In today's rapidly changing environments, the ability to learn from mistakes and readily accept critical feedback is mission-critical for organizations striving to continuously improve performance and stay competitive. However, deeply ingrained cognitive biases pose significant psychological barriers inhibiting individuals and teams from being receptive to negative feedback, even when intended constructively. While awareness of these tendencies is important, proactively addressing them requires innovative solutions beyond traditional performance management approaches. Applied AI and data-driven decision making hold promise in helping overcome many human biases that skew perception and interpretation of feedback by processing multiple perspectives aggregated over time.
Early adopters across industries are demonstrating the value AI can bring by streamlining processes, surfacing consistent themes that may otherwise be missed, and delivering feedback in a depersonalized, evidence-based manner to cultivate a psychologically safer culture of learning. Of course, these technologies must be implemented judiciously alongside comprehensive training, oversight, and respect for privacy/ethics to maximize benefits while minimizing risks or unintended consequences. With continued responsible development and application, AI shows great potential to enhance feedback processes and make organizations more receptive to high-quality criticism as a driver of long-term growth and success. There remains ample opportunity for further research and refinement of these emerging approaches. Overall, strategic use of AI could go far in overcoming innate human resistance to negative input and cultivating a performance environment where feedback is genuinely embraced as a gift rather than a threat.
References
Bommasani, R., Hudson, D. A., Adel, T., Brennan, T., Biega, A. J., Dieterich, S., ... & Fei-Fei, L. (2021). On the opportunities and risks of foundation models. arXiv preprint arXiv:2108.07258.
Bouchet, F., Harley, J. M., & Azevedo, R. (2019). What is the role of an artificial intelligence teaching assistant? Analysis of students' emotions in Anthropic's Constitutional AI. In Proceedings of the Fourth Annual ACM Conference on Learning @ Scale (pp. 1-4).
Bradley, G. W. (1978). Self-serving biases in the attribution process: A reexamination of the fact or fiction question. Journal of Personality and Social psychology, 36(1), 56.
Brynjolfsson, E., & McAfee, A. (2017). The business of artificial intelligence. Harvard Business Review, 95(7), 3-11.
Dietvorst, B. J., Simmons, J. P., & Massey, C. (2015). Algorithm aversion: People erroneously avoid algorithms after seeing them err. Journal of Experimental Psychology: General, 144(1), 114.
Economist. (2018, January 13). How AI is helping to coach salespeople. The Economist. https://www.economist.com/business/2018/01/13/how-ai-is-helping-to-coach-salespeople
Jirak, D., Menold, N., & Sharma, A. (2015). Anonymous feedback to enhance teaching effectiveness and student learning outcomes. Currents in Pharmacy Teaching and Learning, 7(6), 727-733.
Kluger, A. N., & DeNisi, A. (1996). The effects of feedback interventions on performance: A historical review, a meta-analysis, and a preliminary feedback intervention theory. Psychological bulletin, 119(2), 254.
Lai, C. H., & Calvo, R. A. (2015). Effectiveness of digital psych feedback technologies: A systematic review. Computers & Education, 88, 61-67.
Mullaney, T. (2021). Shifting to continuous performance management and feedback with Anthropic PBC. Harvard Business Review. https://hbr.org/2021/05/shifting-to-continuous-performance-management-and-feedback-with-anthropic-pbc
Nickerson, R. S. (1998). Confirmation bias: A ubiquitous phenomenon in many guises. Review of general psychology, 2(2), 175-220.
Steele, C. M. (1988). The psychology of self-affirmation: Sustaining the integrity of the self. In Advances in experimental social psychology (Vol. 21, pp. 261-302). Academic Press.
Tversky, A., & Kahneman, D. (1974). Judgment under uncertainty: Heuristics and biases. Science, 185(4157), 1124-1131.
Jonathan H. Westover, PhD is Chief Academic & Learning Officer (HCI Academy); Chair/Professor, Organizational Leadership (UVU); OD Consultant (Human Capital Innovations). Read Jonathan Westover's executive profile here.
Suggested Citation: Westover, J. H. (2024). How AI Can Make Us More Receptive to Negative Feedback. Human Capital Leadership Review, 12(1). doi.org/10.70175/hclreview.2020.12.1.9