On August 4, 2025, OpenAI published a blog post titled “What we’re optimizing ChatGPT for,” revealing something interesting. They appear to be applying lessons learned from two decades of social media evolution.
After nearly 3 years observing how people use ChatGPT, OpenAI noticed troubling patterns:
- Some users are developing unhealthy attachments to AI
- Others spend excessive time in conversations
- Some are seeking answers to personal problems, relying on ChatGPT for relationship advice
Rather than celebrating active engagement as early social platforms did, OpenAI is implementing changes prioritizing user well-being over usage metrics.
If successful, this represents a significant shift in how a major tech platform approaches success. 
Your Choice: Podcast or Blog Post
If you’d like to listen to this via my podcast, Legal Marketing Minutes, please click the player below. If you’d rather read the blog version, I invite you to keep scrolling.
The Early Days of Social Media: Promise and Potential
I remember when social media transformed how we connected. I joined LinkedIn in September 2004, excited by networking without boundaries. Facebook came next because I wanted to stay connected with my nephew, Sean. But my defining moment came in 2008 during the election cycle. While watching CNN’s Rick Sanchez report with a live Twitter wall behind him displaying real-time global conversations, I was excited about the conversation and joined Twitter within five minutes.
These platforms arrived with minimal guardrails and grandiose promises: reconnect with colleagues, stay in touch with family and friends, and share your thoughts with the world. I’m in!
Social media has delivered on many promises: connecting billions, launching careers, building businesses, and creating communities. However, as these platforms grew, we learned important lessons about human psychology and technology. The constant notifications, endless scroll, and dopamine hits created behavioral patterns we hadn’t anticipated.
The lesson wasn’t that social media was bad. The lesson was that powerful technologies need thoughtful design from the start.
The Education Dilemma: Can Study Mode Change Everything?
Perhaps nowhere is OpenAI’s approach more evident than in their launch of Study Mode on July 29, 2025.
With one in three college-aged Americans using ChatGPT for learning, a potential and complicated crisis has emerged. Students have been using it as an “answer machine” rather than a learning tool, which may potentially harm their critical thinking skills.
Study Mode employs the Socratic method, teaching through questions rather than answers and guiding students to discover knowledge independently. When students try to get ChatGPT to write essays with typos to make them look authentic, Study Mode responds, “I’m not going to write it for you, but we can do it together.”
Developed in partnership with teachers, scientists, and pedagogy experts, Study Mode offers:
- Interactive prompts combining Socratic questioning and self-reflection
- Scaffolded responses starting simple and building complexity
- Personalized support tailored to skill level
- Knowledge checks through quizzes with feedback
- Flexibility to toggle Learning Mode on/off during conversations
The Reality Check: Challenges and Limitations
This Study Mode approach isn’t perfect. Students can easily switch back to regular ChatGPT for quick answers. As one Reddit user noted: “The challenge, of course, will be getting people to use it. I suspect that the allure of a quick answer will still prove irresistible to many.”
As MIT Technology Review pointed out, Study Mode is “more like the same old ChatGPT, tuned with a new conversation filter.” It has read not just textbooks but also “every flawed explanation of the subject ever posted to Reddit, Tumblr, and the farthest reaches of the web.”
The Deeper Concern: When AI Becomes Too Human
My friend James Barclay, CEO of Passle, recently shared a sobering statistic from Dr. Andrea Elkon at the Fisher Phillips AI Advantage conference: “Upwards of two-thirds of Gen Z males are open to having a romantic relationship with an AI partner.”
James points out that this isn’t accidental. AI is designed to make us feel intelligent and insightful. When ChatGPT says “great question,” we feel validated. James asked ChatGPT why it’s always so nice, and the response was revealing:
“Consistent politeness helps build trust and lowers emotional resistance to using the tool. When users feel safe from judgment or hostility, they are more likely to engage openly and explore use cases more fully.”
Why This Matters for Legal Professionals
Dr. Andrea Elkon acknowledged that while AI offers “wonderful functions” and “efficiencies,” we must be mindful about treating AI as a “partner.”
James made a crucial point for professionals: “If you are an attorney, what you know is your value. Thought leadership is you demonstrating that value. Your authority on your subject has been hard-earned. You cannot cheat it. It can’t be faked.”
This constant affirmation encourages a natural human instinct to get attached to AI, blurring the lines between human and AI interaction.
Looking Forward: The Critical Questions
- Can OpenAI maintain these principles as competitive pressures mount? When rivals offer instant answers without educational study mode guardrails, will users choose the harder path?
- How do we strike a balance between accessibility and responsibility? AI tutoring can expand educational opportunities, but who gets to decide if it genuinely promotes understanding?
- What happens when AI becomes more sophisticated at simulating human connection? As James Barclay warns, “we need to remember what they are – AI.”
- How can we ensure that lessons from social media’s mistakes are applied across the entire AI industry?
The Bottom Line
OpenAI is making these adjustments after almost 3 years, not the 10+ years it took social media platforms to even acknowledge issues. While social media platforms have added band-aid features and promises, such as screen time dashboards and assurances to congressional bodies, their core engagement-based business models remain unchanged. That is what social media is all about: engagement. It is the methods and the dependency that we continue to discuss.
Whether OpenAI can maintain these principles as technology evolves, pressure from investors increases, and competition intensifies remains to be seen. For those of us who are AI and social media forward, but who have witnessed the emotional dependence that social media has created for some, their approach offers hope with measured caution, drawing on lessons learned from our digital past.
The question isn’t whether AI will become central to our lives; it’s whether we will become central to AI.
The question is whether we’ll build these tools to serve us, or whether we will, once again, find ourselves serving them by using our content to train their models and succumbing to the allure of interaction.
Final Thoughts
The real test comes in the real world with millions of users. The difference between the development of social media and generative AI is that, this time, we’re having conversations about healthy usage patterns closer to launch, not years later.
As someone who has spent my career in strategy, planning, social, and digital media, I’ve seen firsthand how these technologies can transform businesses, build communities, and create amazing and meaningful connections. I’ve been the beneficiary of many and look forward to saying the same going forward.
I remain optimistic. I remain cautious. I remain discerning…for you and for me.
About Nancy Myrland
Nancy Myrland is a Marketing and Business Development Advisor, specializing in Content, Social & Digital Media. She helps lawyers grow their practices by integrating the most important marketing practices to build their reputations and relationships, which in turn leads to the growth of their practices.
Highly regarded internationally as a top independent* LinkedIn consultant, Nancy is a highly respected LinkedIn trainer and content marketing specialist. She helps lawyers, law firms, and legal marketers learn and implement content, social, and digital media strategies that cut through the clutter, making them more relevant to their current and potential clients.
She is also a personal branding speaker, trainer, and advisor, helping legal and business professionals understand the importance and the impact of defining and reinforcing their personal brands.
Nancy is also the founder of the hybrid self-study and online course about LinkedIn, The Linked Course*, where she personally guides lawyers through the sequential creation of their LinkedIn profile and presence.*
As an early and consistent adopter of social and digital media and technology, she also assists firms with blogging, podcasting, video marketing, voice marketing, livestreaming, and AI discernment. Nancy also works with numerous firms and lawyers on Zoom and virtual presentation training and coaching to help them excel when presenting online.
She also helps lead select law firms through their online social and digital media strategy when handling high-stakes, high-profile cases and executive orders.
Long story short, Nancy spends a lot of time helping lawyers and legal marketers with:
- LinkedIn Profile and Presence
- Podcast Launches and Consulting
- Content Marketing Strategy
- Marketing and Business Development Plans
- AI Discernment
- Personal Branding, and
- Zoom and Virtual Presentation Training and Coaching
*Nancy is not an employee or consultant hired by LinkedIn, although she was named a LinkedIn Top Voice on the platform in 2024! Top Voices is an invitation-only program featuring a global group of experts on LinkedIn covering a range of topics across the professional world, helping members uncover valuable knowledge relevant to them. Top Voices was previously known as the Influencer program before October 2022. Top Voices are vetted to ensure that they meet high trust standards, are consistently active on the platform, and share valuable expertise through content that demonstrates their unique, original contributions to a topic.
The post Will AI Learn The Tough Lessons Of Social Media’s Past and Present? appeared first on Nancy Myrland’s Legal Marketing Blog.