What is Cognitive Surrender?

Artificial intelligence (AI) is making more and more tasks in our daily lives easier: it writes emails for us, summarizes long texts, compares options, and sometimes suggests what we should think or how we should decide. This support can save time by reducing our mental load. However, a significant risk arises when AI moves beyond being merely a tool that assists our thought processes and instead completely takes over the responsibility of thinking. Over time, a person may come to rely on the algorithm not only for the task at hand but also for the processes of questioning, evaluating, verifying, and decision-making. In the literature, this phenomenon is referred to as “cognitive surrender.”

Cognitive surrender refers to the gradual transfer of a person’s responsibility for making logical inferences, critically evaluating information, making ethical decisions, and monitoring their own thought processes to artificial intelligence systems. The fundamental issue here is not merely seeking help from artificial intelligence because people have long been delegating certain cognitive tasks to technology through tools such as calculators, navigation systems, or search engines. What distinguishes cognitive surrender from ordinary cognitive task delegation is that the individual entrusts the system not only with the execution of the task but also with the authority to evaluate the accuracy and appropriateness of the process. In this therapy sketch, we will take a closer look at the term “cognitive surrender”.

What is cognitive surrender?

Imagine you’ve turned on your GPS in a city you’ve never been to before. After a while, you stop paying attention to which streets you’re taking and start automatically following the route shown on the screen. There’s no problem with this as long as the navigation system takes you to the right place. But if you keep driving without looking around even when it directs you down the wrong street, you’re no longer just relying on a tool; you’re handing over control of the navigation process to it. Cognitive surrender describes a situation similar to this in our relationship with AI.

The concept of “cognitive surrender” is defined within the Tri-System Theory of Cognition, developed by Steven D. Shaw and Gideon Nave. The concept is systematically presented in the authors’ work titled “Thinking Fast, Slow, and Artificial: How AI Is Reshaping Human Reasoning and the Rise of Cognitive Surrender” (Shaw & Gideon, 2026).

According to this theory of cognition, traditional dual-process theories explain human thinking through two basic systems:

System 1: A form of thinking that is fast, automatic, intuitive, and requires relatively little effort.
System 2: A form of thinking that is slower, analytical, controlled, and requires effort.

However, according to Shaw and Nave, this model falls short in explaining decision-making processes in the age of AI. This is because thinking no longer takes place solely within the human mind. People rely on external algorithmic systems to find the answer to a problem, compose a text, evaluate options, and even make a decision (Shaw & Gideon, 2026)..

System 3

For this reason, researchers add System 3, that is, “artificial cognition.” System 3 is an algorithmic cognitive system that operates outside the human brain but directly influences human thoughts and behaviors (Shaw & Gideon, 2026). System 3 has four key characteristics:

It is external: It operates not within the human nervous system, but within artificial intelligence models, cloud systems, and algorithmic infrastructures.
It is automated: It performs tasks such as pattern recognition, prediction, summarization, and synthesis through algorithms.
It is data-driven: The results it produces are based on large datasets; therefore, it may also carry the shortcomings and biases of that data.
It is dynamic: It responds to user inputs in real time and interacts with the human thought process.

This system can sometimes support human thinking, sometimes prompt humans to reevaluate their intuitions, and sometimes completely replace the thought process. In this picture, Cognitive surrender is when AI replaces human thinking (Shaw & Gideon, 2026).

We can say that there are three fundamental components of cognitive surrender: the delegation of judgment, cognitive effort, and epistemic responsibility to artificial intelligence (Shaw & Gideon, 2026). For example, let’s consider a college student who uses AI to prepare an assignment. When the student adopts the AI’s perspective as their own without evaluating alternative explanations, they have delegated their judgment. When they incorporate the text into their work without examining the evidence supporting that perspective or comparing it with other sources, they have delegated their cognitive effort. And when they fail to verify the accuracy of the information and assume, “If AI wrote it, it must be true,” they relinquish epistemic responsibility to the system. In this way, they have handed over to AI not only the task of writing the assignment but also the processes of deciding what to believe, questioning information, and taking responsibility for the accuracy of the information they use.

Cognitive Surrender and Mental Health

When we ask AI to write an email, summarize a complex topic, or evaluate our options before making a difficult decision, we are offloading part of our mental load. This offloading is not always negative. AI can organize our thoughts, bring to light options we hadn’t considered, and provide the cognitive support we need under intense stress. However, a different psychological picture emerges when the system that helps us begins, over time, to replace our own thinking and self-regulation processes. Research suggests this distinction may be particularly significant for metacognition, attention, memory, motivation, coping skills, and self-perception.

Metacognitive Awareness: Metacognition is the ability to reflect on one’s own thought processes. AI systems typically generate responses that appear fast and fluent. However, a fluent answer does not mean it is correct. Nevertheless, people may perceive information presented in a clear, organized manner as more reliable. For this reason, AI made answers can weaken the metacognitive cues that would normally trigger skepticism and reevaluation (Jose et al., 2025).
Usage patterns resembling addiction: Because artificial intelligence is fast and accessible, it can reinforce repetitive usage. A person may begin to turn to the system not only for difficult tasks but even for minor issues they could resolve on their own. This could lead to compulsive use and an inability to give up the convenience (Kim et al., 2025).
Motivation and Self-Efficacy: Completing a task through our own efforts not only helps us achieve a result; it also fosters a sense of self-efficacy. When AI simplifies a difficult task, it can reduce the anxiety that prevents a person from getting started. However, if AI consistently completes a significant portion of the task, the person may struggle to distinguish how much of the success is attributable to their own skills. As a result, both the person’s realistic assessment of their own skills may be distorted, and their confidence in their ability to perform the same task without AI may diminish (Zhu et al., 2026).
Supporting or Externalizing Coping Skills: AI can offer significant opportunities in terms of coping. Its immediate accessibility, anonymity, and ability to offer simple suggestions for regulating emotions can reduce barriers to seeking mental health support. For someone experiencing intense stress, suggesting breathing exercises, helping them put their thoughts into writing, or directing them to professional help resources can be beneficial in the short term. However, a different risk emerges when AI begins to constantly label and interpret a person’s emotions on their behalf (Zheng et al., 2025)
Excessive Monitoring and Anxiety: AI-powered apps can continuously track a person’s mood, sleep patterns, productivity, or stress levels. This information can make it easier for a person to recognize their patterns and seek support when needed. However, being constantly scored and receiving feedback on every change can, for some people, increase self-monitoring and anxiety rather than providing relief.

For this reason, when using AI, it is important to consider not only the conveniences it offers but also its effects on mental health and cognitive autonomy. If a person finds it difficult to trust their own decisions, notices that thinking or taking action without AI is becoming increasingly difficult, or realizes that its use is negatively affecting their cognition, mood, and behaviors, they can seek support from a mental health professional.

Take aways

Cognitive surrender involves entrusting the system with the responsibility for judgment, questioning, and verification.
When AI supports our thought process, it can reduce our mental load; however, when it replaces thinking, it can impact our metacognition, motivation, and self-efficacy.
If thinking or making decisions without AI is becoming increasingly difficult, and if this situation is affecting our mood or behavior, we may seek support from a mental health professional.

References

Shaw, Steven D., and Gideon Nave. “Thinking—fast, slow, and artificial: How AI is reshaping human reasoning and the rise of cognitive surrender.” Available at SSRN 6097646 (2026).

Zhu, Q., Li, X., Dong, Y., Chang, P., & Fan, M. (2026). Not all cognitive offloading is equal: distinguishing dependent and autonomous offloading to generative AI. Frontiers in Psychology, 17. https://doi.org/10.3389/fpsyg.2026.1878629

Zeng, R., Vance, D., Joseph, J., Chirayath, G., & Premamalini, K. (2025). Cognitive offloading or cognitive overload? How AI alters the mental architecture of coping. Frontiers in Psychology, 16. https://doi.org/10.3389/fpsyg.2025.1699320

Jose, B., Joseph, D., Mohan, V., Alexander, E., Varghese, S. K., & Roy, A. (2025). Outsourcing cognition: the psychological costs of AI-era convenience. Frontiers in Psychology, 16. https://doi.org/10.3389/fpsyg.2025.1645237

Van Zyl, L. E. (2026). The competence paradox: when psychologists overestimate their understanding of Artificial Intelligence. AI & SOCIETY, 41, 4639 – 4653. https://doi.org/10.1007/s00146-025-02814-9

Kim, T. W., Usman, U., Garvey, A., & Duhachek, A. (2025). From algorithm aversion to AI dependence: Deskilling, upskilling, and emerging addictions in the GenAI age. Consumer Psychology Review. https://doi.org/10.1002/arcp.70008

Book Now

Find a Specialist

At Roamers Therapy, our psychotherapists are here to support you through anxiety, depression, trauma, and relationship issues, race-ethnicity issues, LGBTQIA+ issues, ADHD, Autism, or any challenges you encounter. Our psychotherapists are trained in Cognitive Behavioral Therapy, Dialectical Behavioral Therapy, Psychodynamic Therapy, Acceptance and Commitment Therapy, Person-Centered Therapy, and Gottman Therapy.

Whether you’re seeking guidance on a specific issue or need help navigating difficult emotions, we’re ready to assist you every step of the way.

Contact us today to learn more about our services and schedule a session with our mental health professionals to begin your healing journey. To get started with therapy, you can fill out our inquiry form. You may use this form whether you plan to use insurance benefits, self-pay, or out-of-network benefits, and to receive details about your out-of-pocket costs and our therapists’ availability. Once your appointment is confirmed, you’ll receive intake documents to complete before your first session.

This page is also part of the Roamers Therapy Glossary; a collection of mental-health-related definitions that are written by our therapists.

While our offices are currently located at the South Loop neighborhood of Downtown Chicago and Lakeview on Chicago’s North Side, Illinois, we also welcome and serve clients for online therapy from anywhere in Illinois and Washington, D.C. Clients from the Chicagoland area may choose in-office or online therapy and usually commute from surrounding areas such as River North, West Loop, Gold Coast, Old Town, Lincoln Park, Rogers Park, Logan Square, Pilsen, Bridgeport, Little Village, Bronzeville, South Shore, Hyde Park, Back of the Yards, Wicker Park, Bucktown and many more. You can visit our contact page to access detailed information on our office location.