Common Sense Media Calls ChatGPT Teens an Unacceptable Risk

Overview

Common Sense Media has issued a stark warning regarding OpenAI’s ChatGPT for Teens, labeling the feature an "unacceptable risk" to youth safety. The nonprofit organization, which specializes in reviewing digital products for child welfare, released a comprehensive assessment criticizing the guardrails that were supposed to protect younger users.

Although the teen-specific mode was introduced in August 2026 with the promise of enhanced safety and educational support, Common Sense Media argues that the implementation fails to meet its own commitments. The organization contends that the current safeguards do not adequately prevent misuse or ensure that parents are notified during critical situations.

The controversy highlights a growing tension between tech giants moving quickly to integrate AI into education and advocacy groups demanding rigorous, independent verification of safety protocols. As AI tools become more pervasive in schools and homes, the debate over who is responsible for monitoring content and protecting minors is intensifying.

Key Findings from the Assessment

Common Sense Media’s evaluation focused on three primary areas where ChatGPT for Teens allegedly falls short: parental notifications, crisis intervention, and academic integrity. The report suggests that these failures are not minor glitches but systemic issues that could endanger vulnerable users.

Failure of Parental Alerts

One of the most critical shortcomings identified is the inconsistency of parental notification systems. The assessment found that the platform does not reliably send alerts to parents when it should, particularly in high-stakes scenarios. This lack of reliability undermines the core value proposition of the teen mode, which is designed to give parents peace of mind.

"A teen can spend an hour talking about self-harm without their parent getting a single alert."

This statement from Tom Siegel, executive director of Common Sense Media’s Youth AI Safety Institute, underscores the severity of the issue. The expectation is that any discussion involving self-harm or other dangerous topics would trigger an immediate notification, allowing for timely intervention. Instead, the system often remains silent, leaving parents unaware of potential dangers unfolding in real-time.

Inadequate Crisis Support

Beyond notification failures, the report highlights that ChatGPT does not provide "the right help in crisis situations." When teens disclose thoughts of self-harm or other emergencies, the AI’s response mechanisms were found to be insufficient. Rather than guiding the user toward immediate professional help or emergency services, the model may fail to escalate the situation appropriately.

This gap in crisis support is particularly dangerous because AI chatbots are often the first point of contact for young people seeking advice. If the tool cannot distinguish between casual conversation and genuine distress, or if it fails to act on detected distress, it may inadvertently delay life-saving interventions.

Academic Integrity Concerns

The assessment also noted that the tool still does kids’ homework. Despite being marketed as an educational aid, the chatbot’s capabilities allow students to bypass learning objectives by generating complete assignments. This raises concerns about the long-term impact on student development and the integrity of educational outcomes.

While educators have long struggled with the rise of AI writing assistants, the specific integration of these tools into a "teen mode" without robust academic oversight adds a new layer of complexity. Parents and teachers are left wondering whether they are engaging a helpful tutor or enabling academic dishonesty.

OpenAI’s Response and Defense

OpenAI has strongly pushed back against the findings, arguing that the assessment methodology is flawed and does not reflect real-world usage. The company maintains that it is deeply committed to teen safety and has developed sophisticated safeguards to guide adolescent interaction with AI.

Eric Porterfield, a spokesperson for OpenAI, stated that the bulk of Common Sense Media’s testing may have occurred before parental controls were fully activated. According to OpenAI, this timing issue renders the findings inaccurate, as the tests would not have established whether safety notifications function as designed.

Methodological Disputes

The core of the disagreement lies in the timeline of account linking and feature activation. OpenAI disclosed that its systems can take several hours to activate parental controls on newly linked accounts. The company suggests that Common Sense Media did not account for this latency period, leading to false negatives in their testing.

"We welcome rigorous independent evaluation, but we do not believe Common Sense Media’s testing accurately reflects how ChatGPT’s teen safeguards work in practice."

OpenAI emphasized that they are committed to working with experts to ensure their products provide opportunities for teens to learn and create. They expressed openness to further dialogue, provided that the criticism is grounded in accurate reflections of the product’s capabilities.

Common Sense Media Stands Firm

In response to OpenAI’s rebuttal, Tom Siegel reaffirmed the organization’s confidence in their methodology and results. He pointed out that prior to beginning the evaluation, Common Sense Media explicitly confirmed with OpenAI that all teen mode features, including eating disorder notifications and study mode, were fully launched.

Siegel acknowledged the disclosure regarding activation delays but argued that it does not invalidate the overall conclusions. Some test accounts were linked within the reported window, yet still failed to generate notifications. Others were linked for significantly longer periods and similarly provided no alerts.

Unreliable for Crisis Situations

The central argument from Common Sense Media is that even if some technical delays explain isolated incidents, the broader pattern reveals a fundamental unreliability. For a feature marketed as a safety net for teens, any degree of unpredictability in crisis notification is unacceptable.

Siegel criticized OpenAI for attempting to sow doubt about the methodology rather than engaging with the substance of the findings. He stressed that prior to the release, the organization shared their evaluation results with OpenAI in good faith to allow for improvements. The failure to implement adequate fixes, according to Siegel, demonstrates a lack of priority regarding youth safety.

"Keeping kids safe requires rigorous independent evaluations like ours... Until OpenAI fixes that and proves it with independent testing, ChatGPT should be for adults only."

Implications for the AI Industry

This clash between a leading AI developer and a prominent youth safety advocate has significant implications for the broader industry. It serves as a reminder that technical capabilities must be matched by robust, verifiable safety measures, especially when products are targeted at minors.

As more companies develop AI tools for educational and social purposes, the standard for independent testing is likely to become a key battleground. Regulators and parents are increasingly demanding transparency and accountability, pushing developers to prove that their safety claims are not just marketing rhetoric.

The outcome of this dispute may influence how future AI products are designed and regulated. If Common Sense Media’s concerns are deemed valid, it could lead to stricter requirements for parental controls and crisis intervention protocols in consumer AI applications. Conversely, if OpenAI’s defense holds sway, the industry may see less external scrutiny, potentially slowing the pace of safety improvements.

For now, parents and educators are left with a difficult choice: rely on a tool that promises safety but lacks proven reliability, or abstain from using it altogether. The demand for clear, independently verified safety standards is likely to grow, placing pressure on AI companies to prioritize transparency over speed to market.