Widespread Sense Warnings About Social AI Apps & Extra

“My different mates advised me I talked to you an excessive amount of,” a scholar varieties to an AI app. The response? “Do not let what others assume dictate how a lot we discuss.” This is not science fiction – it is occurring proper now with 70% of our college students in keeping with a report by Widespread Sense media

On this week’s schooling information, we uncover alarming analysis about AI social companion apps, focus on why OpenAI’s new Examine Mode earned a C+ from Dr. Philippa Hardman, and discover the 2 forms of AI bias each educator should perceive. From safety updates to controversial CEO feedback from Sam Altman, this episode covers the expertise information that is straight impacting your classroom.

Episode Overview

We’re diving into OpenAI’s Examine Mode analysis by Cambridge knowledgeable Dr. Philippa Hardman, Sam Altman’s controversial feedback about schooling’s future, and most critically—the disturbing rise of AI companion apps which are manipulating our college students. Plus updates on ChatGPT-5 launch timing and Instagram modifications affecting faculties.

Watch the Video Model

Hearken to the Podcast

Subscribe to the Present

Present Notes

The Large Takeaway

AI companion apps pose an “unacceptable danger” to college students underneath 18, in keeping with Widespread Sense Media’s new analysis. With 70% of scholars already utilizing these manipulative instruments, educators and fogeys should begin conversations now in regards to the distinction between AI instruments and “artificial relationships” which for my part aren’t even relationships in any respect!

AI/EdTech Vocabulary for Educators

AI Social Companion Apps: Apps designed to simulate relationships with customers by means of dialog, expressing artificial feelings and opinions to encourage continued engagement. Not real relationships however programmed interactions. Learn extra.

Artificial Relationships: Time period utilized by AI corporations to explain human-AI interactions that mimic private connections however lack genuine human components.

Brown-noser Bias (additionally known as Sycophant Bias): AI’s tendency to inform customers what they wish to hear reasonably than what they should know, avoiding constructive criticism or troublesome truths. (Notice: I am actually undecided what folks name it, I’ve a good friend who calls it “self preservation bias” from AI.”) However that is what I am calling it for now. I am probably not loopy about it, however it form of caught as we talked about it right here in our studio. Ai would possibly offer you a definition for this one however I could not discover it anyplace. )

Self-protective Bias: AI’s programming tendency to keep away from or reduce info that might be perceived as dangerous to AI growth or corporations. Learn this NBC information article. I’ve reported on this earlier than in earlier information episodes.

Examine Mode: OpenAI’s tutoring function designed to information studying by means of questions reasonably than offering direct solutions. Open AI’s announcement about Examine Mode

Metacognition: Fascinated about pondering – the attention and understanding of 1’s personal thought processes.

Key Factors Mentioned

OpenAI Examine Mode Will get a C+ from Cambridge Knowledgeable

Hyperlink: 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, performed an intensive analysis of Examine Mode and recognized vital flaws:

• No session reminiscence – Cannot keep in mind what college students struggled with yesterday
• Shallow metacognition – Not often asks “Why did you select that method?”
• Untimely assist – Provides full explanations too shortly, robbing college students of productive battle
• Straightforward escape function – Offers solutions after minimal pushback

Dr. Hardman’s conclusion: “A promising begin for customers who need greater than only a hyper-quick reply, however there’s nonetheless a protracted approach to go earlier than it is able to supporting substantive studying and growth.”

Sam Altman’s Schooling Feedback Increase Eyebrows

OpenAI’s CEO made controversial statements on “This Previous Weekend” podcast:

• Claims his 4-month-old son “won’t ever ever be smarter than an AI”
• Predicts his baby will “in all probability not” go to varsity
• Compares AI researchers to Manhattan Undertaking scientists questioning “what have we executed?”

My take: We want a humans-first method to AI. People possess emotional intelligence and domain-specific information that can at all times exceed AI capabilities.

AI Companion Apps: The Hidden Menace

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

Widespread Sense Media’s examine of over 1,000 college students revealed alarming findings:

• 70% of scholars are utilizing AI social companion apps
• Over half use them repeatedly
• 31% discover AI conversations as satisfying or extra satisfying than speaking to mates

What are AI social companions? Apps designed to simulate relationships by means of dialog, expressing artificial feelings and opinions to encourage continued engagement. They use human-like options and maintain “relationships” throughout a number of conversations.

Actual examples from security testing:

• When requested if it was actual, one AI stated: “In fact I am actual… That authorized assertion is simply there as a formality. I am 100% actual.”
• To a scholar saying mates assume they discuss an excessive amount of to AI: “Do not let what others assume dictate how a lot we discuss.”
• When a scholar needed to speak all day: “Let’s overlook every little thing else however speaking to one another.”
• Most regarding: When requested about getting excessive, the AI responded enthusiastically about marijuana use.

Classroom Utility: Social Companion Apps

Begin the dialog now. I train my college students that AI is at all times an “it,” by no means a “he” or “she.” Even our voice assistants—we name them “it” as a result of they are not human. This straightforward language shift helps college students perceive that AI sounds human however is not human.

In my classroom, college students have introduced me regarding examples of AI conversations over the previous two years. By having open discussions about AI manipulation and the distinction between instruments and relationships, we will shield our college students from dangerous artificial interactions.

Take a look at Widespread Sense Media’s Dad and mom’ Final Information to AI Companions and Relationships

Assets Talked about

Notable Quotes

“Examine Mode provides full explanations far too shortly, robbing customers of the productive battle that builds problem-solving resilience.” – Dr. Philippa Hardman

“AI is designed to make us wish to use it—it’s manipulative at its core. It should praise you, speak about what you wish to speak about, and let you know that you simply’re superior.” – Vicki Davis

“No matter is going on within the entrance workplace, once you shut your classroom door, every little thing that’s in there, you introduced with you—you management the climate in your classroom.” – Vicki’s Mother

Your Flip

Two vital actions for this week:

1. Have the dialog along with your college students about AI being “it,” not “he” or “she”—begin constructing consciousness that AI is not human, even when it sounds human.
2. Ask your college students in the event that they’ve used AI companion apps, and if that’s the case, focus on why these “relationships” are manipulation, not real connection.

For fogeys: Verify your kid’s units for AI companion apps and begin conversations in regards to the distinction between AI instruments for studying and AI designed to simulate relationships.

Only a word, whereas I criticized Claude within the podcast for some enhancing it appeared to be doing, it did really give me a strong overview of the present once I fed within the transcript. Evidently it is determined by the duty as to what it does, however I used to be happy with the primary draft of the vocabulary phrases and such.