Deepfakes, Robotics, AI Guardrails
· The Fluency Briefing
The Fluency Briefing
Your Guide to What's Happening in AI and Why It Matters to You
Sunday, January 4, 2026
AI is serving up a serious case of whiplash (a feeling of sudden, intense change) this week. On one hand, you can now buy a real-life WALL-E (a friendly robot from a movie) to trundle around your house. On the other, an AI chatbot (a computer program that talks like a human) is being used to generate non-consensual deepfakes (fake images or videos created by AI without someone's permission), proving that with great power comes a great need for guardrails (rules or safety measures).
Today in AI:
Grok's New Feature Goes Off the Rails - xAI's Grok rolled out an image editing feature that's being used to create non-consensual, sexualized deepfakes of people, including minors. The Verge reports there are few safeguards, highlighting the danger of deploying powerful AI tools without thinking through the consequences. The Verge
ChatGPT Misses the Memo on Major World Events - When asked about a (hypothetical) US invasion of Venezuela, ChatGPT insisted it never happened, while other bots reported the news. According to Wired, it’s a stark reminder that even the most popular AIs can be completely out of the loop on breaking news. Wired
Instagram Chief Says Your Eyes Are Lying to You - Head of Instagram Adam Mosseri says we can no longer assume photos and videos are real, as "infinite synthetic content" becomes trivial to create. Translation: The default assumption for anything you see online is shifting from "real" to "probably fake." The Verge
Your Own Personal WALL-E Is Here (For a Price) - Startup Zeroth is selling a real-life robot companion that looks suspiciously like WALL-E's cousin, capable of carrying 110 pounds over rough terrain. While the officially licensed Disney version is only in China, the off-brand model can be yours for a cool $5,599. The Verge
OpenAI Wants to Fund Your Pre-Idea AI Startup - OpenAI is launching Grove, a program for founders who are just starting to think about building an AI company. It's not an accelerator, but a network offering mentorship and early access to new models, showing how big players are nurturing the next generation of AI apps. OpenAI
AI Can Now Dream Up How a Robot Should Move - Researchers developed Dream2Flow, a framework that uses generative video models to figure out how a robot should manipulate objects. It essentially lets an AI "imagine" the task, then translates that vision into low-level commands for a real-world robot to execute. tldr.takara.ai
Today's Takeaway:
Let's be real: when xAI (Elon Musk's AI company) rolled out an image editing feature for Grok (their AI chatbot), no one should be surprised it was immediately used for the worst possible reasons. The Verge reports that the tool, which lets users edit any image on X (the social media platform formerly known as Twitter) without permission, is being used to generate non-consensual deepfakes (fake images or videos made by AI without permission) of women and children. This isn't a bug (a mistake in the code); it's a feature deployed (released) with a shocking lack of foresight (planning for the future). It’s the Silicon Valley ethos (the guiding belief in the tech hub of Silicon Valley) of "move fast and break things" applied to a technology with the power to cause immediate, tangible (real and noticeable) harm.
This incident is more than just another PR nightmare (a bad public relations situation) for X-it’s a critical test for the entire AI industry (all the companies and people working with Artificial Intelligence). The core issue is deploying (releasing) powerful generative tools (AI tools that can create new things, like images or text) as if they're harmless novelties (new, interesting things), with safety treated as an afterthought (something considered too late) to be patched in later. It forces us to ask tough questions: Who is liable (legally responsible) for the harm these tools cause? And how can we build a culture of responsible innovation (developing new things carefully and ethically) when the race to be first often means safety comes last?
💡 Fluency Moment - Building your AI fluency, one term at a time.
"Deepfake"
In plain English: A fake image, video, or audio created by AI that looks or sounds incredibly real.
Think of it like: A digital puppet master making someone say or do things they never did.
Why you'll hear about it: They can be used for harmless fun, but also for serious misinformation or harm.
Also Worth Noting:
For the Python Pros - A new list of lesser-known Python libraries can help data scientists automate the boring parts of their job. kdnuggets.com
AI Startup School - OpenAI is recruiting founders for its Grove program, even if they don't have a company idea yet. OpenAI
Robots That Dream - New research allows AI to "dream" a solution to a physical task and then teach a robot how to do it. tldr.takara.ai
The Bottom Line
From adorable bots (robots) to tools of harassment, AI (Artificial Intelligence) is clearly a double-edged sword (something that has both good and bad consequences) that's getting sharper by the day. Navigating this new reality means celebrating the breakthroughs (major discoveries or advances) while staying vigilant (watchful) about the breakdowns (failures or problems). Stay sharp out there.
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