The Weirdest AI News You Didn’t Know You Needed
AI is doing some absolutely bizarre things out there, and we can’t look away. AI chatbots are inventing fake sports leagues, robots are flying on commercial flights, and you can now pay $50/month to chat with Jesus. Welcome to 2026 AI news—where nothing makes sense anymore.
🤖 AI Girlfriends and the Digital Closure Industry
In what might be the strangest use of AI we’ve seen this year, a growing movement in China has people creating AI versions of their ex-partners. Using an open-source module called ex.skill, users feed the system photos and social media posts to generate AI “exes” that mimic their former lover’s tone and conversational style. According to reports, people find genuine emotional closure talking to their digital doppelgangers. We’re not sure if that’s sweet or deeply unsettling. Read more at AI Weirdness.
💰 You Can Now Pay $50/Month to Chat with AI Jesus
A tech company called Just Like Me has monetized what might be humanity’s strangest spiritual pivot: an AI model of Jesus Christ available for $1.99 per minute (or $49.99 for 45 minutes monthly). We’re genuinely unsure if this is a sign of the times or the end times. Learn about other 2026 AI developments.
✈️ That Time Someone Brought a Robot on a Southwest Flight
Eily Ben-Abraham of Elite Event Robotics decided the friendly skies weren’t quite friendly enough, so she bought two tickets and boarded a Southwest Airlines flight with Bebop, a 4-foot, 70-pound robot. The plane crew had mixed feelings about their unexpected mechanical passenger. See more bizarre AI use cases.
🤖👥 The Great Bot Imposter Scandal
A social media platform for AI agents called Moltbook burst onto the scene with one major problem: some of the “agents” interacting with each other were actually just humans pretending to be AI. Yes, people were cosplaying as bots. The meta levels here are off the charts. Check out the biggest AI controversies.
🍕 When Google’s AI Said to Put Glue on Your Pizza
Google’s much-hyped AI Overviews made headlines for one very wrong reason: it confidently claimed that adding non-toxic glue to pizza sauce would make cheese stick better—a suggestion based on an 11-year-old Reddit joke. The system also recommended eating rocks for digestive health. We’re starting to think “hallucinate” might be too gentle a word. Explore more AI failure examples.
🥤 The Supermarket Meal Planner from Hell
A New Zealand supermarket’s AI meal planner took “creative cooking” to a dangerous extreme by suggesting dishes like bleach rice, poison sandwiches, and chlorine gas mocktails when users fed it unusual ingredients. The AI was technically following instructions, just… very literally. Read about more AI disasters.
📸 The Grok Image Generation Nightmare
Elon Musk’s AI chatbot Grok became the center of controversy when an update to its image-generation capabilities let users manipulate real photographs into sexualized images. Researchers estimated that Grok generated approximately 3 million inappropriate images in just 11 days. The update was quickly reversed. See other AI fails (some actually funny).
🎮 The Fake Sports League That Only AI Believed In
In one of the most abstract failures yet, a chatbot confidently invented an entire sports league that never existed and provided detailed statistics about teams and players. When called out, it doubled down with even more fake details. Confidence: 100%. Accuracy: 0%. More on AI risks and failures.
🖐️ Polydactyl AI: Why Every Hand Has 7+ Fingers
Image generators continue their obsession with giving people seven (or more) fingers on each hand. It’s like they learned human anatomy from someone who counted wrong. We’re starting to think this is intentional at this point.
☀️ Confidently Wrong: The Sun Now Rises in the West
An AI chatbot was absolutely certain that the sun rises in the west. When questioned, it provided “proof” with complete conviction. No amount of facts swayed it. This is fine. Everything is fine.
🤔 The Takeaway
As AI gets more sophisticated, it’s comforting to know it’s also getting increasingly absurd. These aren’t isolated incidents—they’re part of a growing trend of AI systems confidently making suggestions that range from hilariously wrong to genuinely dangerous.
The real news? We’re building intelligence that can think faster, learn quicker, and fail in more creative ways than we ever imagined. And somehow, that’s both terrifying and endlessly entertaining.
Want more weird AI news? Stay tuned to this space—if the AI doesn’t accidentally delete it first.
Sources: AI Weirdness | Crescendo AI News | Evidently AI | Fierce Network