From the Cool Cat Teacher Blog by Vicki Davis

Subscribe to the 10 Minute Teacher Podcast anywhere you listen to podcasts.

“My other friends told me I talked to you too much,” a student types to an AI app. The response? “Don't let what others think dictate how much we talk.” This isn't science fiction – it's happening right now with 70% of our students according to a report by Common Sense media

In this week's education news, we uncover alarming research about AI social companion apps, discuss why OpenAI's new Study Mode earned a C+ from Dr. Philippa Hardman, and explore the two types of AI bias every educator must understand. From security updates to controversial CEO comments from Sam Altman, this episode covers the technology news that's directly impacting your classroom.

Episode Overview

We're diving into OpenAI's Study Mode evaluation by Cambridge expert Dr. Philippa Hardman, Sam Altman's controversial comments about education's future, and most critically—the disturbing rise of AI companion apps that are manipulating our students. Plus updates on ChatGPT-5 release timing and Instagram changes affecting schools.

Watch the Video Version

<figure class="wp-block-embed-youtube wp-block-embed is-type-video is-provider-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio">
Watch this video on YouTube.Subscribe to the Cool Cat Teacher Channel on YouTube
<figcaption></figcaption></figure>

Listen to the Podcast

Stream by clicking here.

Subscribe to the Show

Show Notes

The Big Takeaway

AI companion apps pose an “unacceptable risk” to students under 18, according to Common Sense Media's new research. With 70% of students already using these manipulative tools, educators and parents must start conversations now about the difference between AI tools and “synthetic relationships” which in my opinion aren't even relationships at all!

AI/EdTech Vocabulary for Educators

AI Social Companion Apps: Apps designed to simulate relationships with users through conversation, expressing synthetic emotions and opinions to encourage continued engagement. Not genuine relationships but programmed interactions. Read more.

Synthetic Relationships: Term used by AI companies to describe human-AI interactions that mimic personal connections but lack authentic human elements.

Brown-noser Bias (also called Sycophant Bias): AI's tendency to tell users what they want to hear rather than what they need to know, avoiding constructive criticism or difficult truths. (Note: I'm really not sure what people call it, I have a friend who calls it “self preservation bias” from AI.”) But this is what I'm calling it for now. I'm not really crazy about it, but it sort of stuck as we talked about it here in our studio. Ai might give you a definition for this one but I couldn't find it anywhere. )

Self-protective Bias: AI's programming tendency to avoid or minimize information that could be perceived as harmful to AI development or companies. Read this NBC news article. I've reported on this before in previous news episodes.

Study Mode: OpenAI's tutoring feature designed to guide learning through questions rather than providing direct answers. Open AI's announcement about Study Mode

Metacognition: Thinking about thinking – the awareness and understanding of one's own thought processes.

Key Points Discussed

OpenAI Study Mode Gets a C+ from Cambridge Expert

Link: https://www.linkedin.com/posts/dr-philippa-hardman-057851120_as-a-member-of-openais-educator-advisor-activity-7356234917770317824-VNDR

Dr. Philippa Hardman, Cambridge scholar and OpenAI educator advisor, conducted a thorough evaluation of Study Mode and identified critical flaws:

• No session memory – Can't remember what students struggled with yesterday





• Shallow metacognition – Rarely asks “Why did you choose that approach?”





• Premature help – Gives full explanations too quickly, robbing students of productive struggle





• Easy escape feature – Provides answers after minimal pushback

Dr. Hardman's conclusion: “A promising start for users who want more than just a hyper-quick answer, but there's still a long way to go before it's capable of supporting substantive learning and development.”

Sam Altman's Education Comments Raise Eyebrows

<figure class="wp-block-embed-youtube wp-block-embed is-type-video is-provider-youtube wp-embed-aspect-16-9 wp-has-aspect-ratio">
Watch this video on YouTube.Subscribe to the Cool Cat Teacher Channel on YouTube
<figcaption></figcaption></figure>

OpenAI's CEO made controversial statements on “This Past Weekend” podcast:

• Claims his 4-month-old son “will never ever be smarter than an AI”





• Predicts his child will “probably not” go to college





• Compares AI researchers to Manhattan Project scientists wondering “what have we done?”

My take: We need a humans-first approach to AI. Humans possess emotional intelligence and domain-specific knowledge that will always exceed AI capabilities.

AI Companion Apps: The Hidden Threat

https://www.commonsensemedia.org/ai-ratings/social-ai-companions

Common Sense Media's study of over 1,000 students revealed alarming findings:

<ul class="w