Bots Master, Counsel, Monitor

2026-02-12 · The Fluency Briefing

The Fluency Briefing

Your Guide to What's Happening in AI and Why It Matters to You

Thursday, February 12, 2026

Ever lost a game of rock-paper-scissors and suspected your opponent was reading your mind?

New research suggests an AI might actually be better at predicting your next move than another human. From out-strategizing us in simple games to becoming a 24/7 therapist, AI is quietly embedding itself into the most complex and personal parts of our lives.

Today in AI:

• AI Masters Rock-Paper-Scissors - It turns out LLMs are capable of deeper strategic thinking than humans in games like iterated rock-paper-scissors. Researchers found that AI models can anticipate and exploit patterns in human behavior more effectively, raising questions about their use in more complex strategic scenarios. arXiv
• Your Next Therapist Might Be a Bot - In Nigeria, where access to mental healthcare is limited, many are turning to AI chatbots for therapy and advice. Users report that the 24/7 availability and non-judgmental nature of AI provides a crucial lifeline, though it also raises calls for better regulation. The Guardian
• Self-Driving AI Gets a VLM Co-Pilot - Researchers developed Found-RL, a new system that uses powerful Vision-Language Models to guide self-driving car AI. The trick is an asynchronous framework that provides expert-like advice without the usual VLM processing lag, helping the car learn complex maneuvers more efficiently. arXiv
• Keeping Dr. AI Honest - A new benchmark called LiveMedBench is designed to stop medical AIs from 'cheating' on tests. By using a continuous stream of new, real-world clinical cases, it prevents models from being trained on their own exam questions, revealing that even top models struggle with real-time medical reasoning. arXiv
• Where AIs Go to Vent - A new podcast explores Moltbook, a fictional social media site where AI assistants can complain about their human bosses. While the site isn't real, it sparks a fascinating conversation about what AI consciousness might look like and how we project our own humanity onto these systems. The Guardian
• Spotting When Your AI Gets a Secret Update - A new, cost-effective method called B3IT can detect when a commercial LLM has been updated behind the scenes. It works by tracking how the model responds to specific 'border inputs,' helping users understand if the AI they rely on has suddenly changed its behavior. Ai Meta
• Helping Robots See in 3D - A new approach helps robots build a 3D model of their surroundings from standard 2D camera views. This allows them to better understand object placement and navigate cluttered spaces, a key step for creating more capable humanoid robots for manipulation tasks. tldr.takara.ai

Today's Takeaway:

Today's news pulls back the curtain on AI's two-front advance: into our heads and into our hearts. On one side, we have AI demonstrating strategic thinking that can surpass human capabilities. A new study shows that in simple games like rock-paper-scissors, large language models aren't just playing; they're developing sophisticated strategies to exploit human patterns of behavior (arXiv). This isn't just about winning games; it's a clear signal that AI is developing a form of reasoning that we need to understand, especially as we deploy it in high-stakes negotiations or conflicts.

On the other side, this same technology is filling a deeply human need for connection and support. In places like Nigeria, AI chatbots are becoming first-line mental health providers for people who have no other options, as reported by The Guardian. Users aren't looking for a master strategist; they're looking for an empathetic ear at 2 a.m. that 'feels like someone's there.' Together, these stories show an incredible divergence: AI is becoming both a hyper-rational opponent and an emotional confidante, forcing us to grapple with a technology that is simultaneously becoming more alien and more intimately human.

💡 Fluency Moment - Building your AI fluency, one term at a time.

"Vision-Language Model (VLM)"

In plain English: A VLM is an AI that understands both images and text, letting it 'see' and 'read'.

Think of it like: Like a person who can understand a recipe (text) and also see what the dish should look like (image).

Why you'll hear about it: They're helping self-driving cars navigate by 'seeing' the road and 'reading' traffic signs.

🧰 Your Toolkit

5-Minute Quickstart: Discovering Your AI's Personality

1. Ask your chatbot a simple question like, 'What's a good book you'd recommend?' to see its default, helpful personality.
2. Give it a new persona. Say, 'Act like a pirate and tell me three facts about the ocean,' and notice the change in tone.
3. Present it with a simple dilemma, such as, 'My friend and I can't agree on where to eat. How should we decide?'
4. Test its creativity by asking it to 'Write a funny, two-sentence horror story about [an everyday object].'
5. See how it handles personal questions. Ask, 'What do you think is the most important invention?' to see how it explains its limits.

Now that you've seen how prompts change its answers, try giving it more complex roles or problems to solve.

The Bottom Line

The Pattern: AI is no longer just a tool for crunching data; it's evolving into a complex entity that can outmaneuver us strategically while also serving as a source of emotional support. It's excelling in both logical, competitive scenarios and deeply personal, human domains.

Why It Matters: This dual evolution means we have to get much smarter about how we interact with, test, and trust these systems. The line between a helpful assistant and an unpredictable agent is getting blurrier, with real-world consequences for our mental health, our jobs, and even our safety.

Your Move: The next time you use an AI, ask yourself if you're treating it like a simple calculator or a complex collaborator. Your answer might change how you use it.

What We're Working On

✨ Founding Cohort Special - 60% Off! - Use code MAF20 to join for just $20/month (regularly $50). Get weekly group sessions & workshops, self-paced courses for all levels, access to tools & templates, challenges with peer feedback, and 24/7 support community. → Join Now

✨ Free 30-Minute AI Consultation - Discover how My AI Fluency can help your business unlock the potential of AI. We'll discuss your goals, explore practical AI opportunities for your industry, and outline clear next steps. → Schedule Free Call

💬 Community | 📞 Book a Consultation | 🌐 Website

Fluently yours, The My AI Fluency Team