- Horizon AI
- Posts
- These States Use AI More Than Anyone Else in 2026 🤖
These States Use AI More Than Anyone Else in 2026 🤖
Research Repository arXiv Cracks Down on AI-Generated Papers 🤯
Gina Acosta
May 24, 2026
Sponsored by
Welcome to another edition of Horizon AI,
By applying the same deep learning algorithms that power LLMs like ChatGPT to a massive database of marine recordings, scientists have unraveled a hidden structure beneath the sea. They found that sperm whale communication isn't random animal noise but a highly sophisticated linguistic system, structurally identical to the building blocks of human speech.
Let’s jump right in!
Read Time: 4.5’ min
Here's what's new today in the Horizon AI
Chart of the week: The States Where Businesses Use AI Most
ArXiv Will Ban Researchers for a Year Over Unchecked AI-Generated Papers
AI Findings/Resources
AI tools to check out
Video of the week
TOGETHER WITH READDY
Finally, an AI Website You're Not Embarrassed to Ship
Most AI builders get you 80% there. Then you're on your own, having to stitch together a complicated web of tools: hosting, SEO, payments, backend, deployment, and more.
Readdy takes you all the way:
Full website generated from a single prompt
Hosting, SEO, analytics, and payments built in
24/7 AI agent that handles customer inquiries
Go live in one click. No DevOps. No extra tabs.
$15/month. Everything included. And a site you’re actually proud of.
Chart of the week
The States Where Businesses Use AI Most
Colorado and Arizona lead the U.S. in business AI adoption in 2026, while California ranks lower than many expect at 13th nationally.
States with lower adoption rates tend to have economies centered on agriculture, manufacturing, or smaller businesses, where AI deployment has been slower.
Company size matters even more than geography, with large firms adopting AI at far higher rates than small businesses.
AI News
AI IN ACADEMIA
ArXiv Will Ban Researchers for a Year Over Unchecked AI-Generated Papers
Research repository ArXiv is introducing stricter rules against careless AI use in scientific papers, warning that authors could face a one-year ban if they submit work containing obvious unverified AI-generated content.
Details:
ArXiv says papers showing clear signs that authors failed to review AI-generated material will now trigger penalties, including a one-year suspension from submitting to the platform.
Examples of unacceptable content include hallucinated references, fabricated citations, or leftover prompts and comments generated by large language models.
After the suspension, affected researchers will only be allowed to post new papers if those papers are first accepted by a reputable peer-reviewed journal or conference.
Moderators and section chairs will review evidence before issuing penalties, and authors will have the ability to appeal decisions.
The policy does not ban the use of AI tools entirely. Instead, ArXiv says researchers must take full responsibility for everything included in their papers, regardless of whether the content was written by humans or generated by AI.
The move comes amid growing concerns across academia about the rise of AI-generated errors and fake citations appearing in scientific research papers.
AI Findings/Resources
😱 Chinese students are buying GPT-5.4/5.5 and Claude API access from Xianyu/Taobao proxy sellers for almost 96-97% cheaper
👉 Google published a long piece about "Optimizing your website for generative AI features on Google Search"
AI Tools to check out
💥 AI Create: free, open-source AI tools for images, video, audio, and text, all running locally on your device.
📹 Ponder: an agentic AI video editor that collaborates with you to craft world-class stories in minutes.
🔍 Cubic: Al code review platform that helps teams catch bugs and merge pull requests faster 28% faster.
🦾 VideoAI.me: Transform ideas into professional-quality videos with custom human-like actors & realistic voiceovers.
🌶️ Pipali: an AI co-worker on your computer that researches, creates documents, and automates workflows.
Video of the week
Scientists Used AI to Decode Whale Communication
For decades, scientists believed that sperm whales communicated using a simple, fixed menu of roughly 21 distinct click patterns called "kodas". However, running these sounds through advanced artificial intelligence has structurally demolished the idea that humans are the only species on Earth to independently evolve complex language.
The Accidental Breakthrough & Project CETI
The breakthrough began when marine biologist David Gruber was playing recordings of sperm whales in his Harvard office. A cryptographer walking by recognized that the rapid mechanical clicking sounded less like random noise and more like an encrypted Morse code system that had never been analyzed by a machine learning decoder .
Project CETI Launched: In 2020, Project CETI (the Cetacean Translation Initiative) was founded alongside MIT, Harvard, and UC Berkeley. The team deployed underwater hydrophone arrays and advanced biometric suction tags to capture 9,000 whale recordings alongside granular behavioral context like physical movement and social hierarchy.
What the AI Found: An Underwater Phonetic Alphabet
From 21 to 156 Kodas: The AI didn't just find a few new signals, it mapped 156 distinct kodas, complete with internal structure, tempo shifts, and layered click patterns.
Rather than pulling from a static menu of messages, the whales construct their communication dynamically. Similar to how English builds thousands of words from 44 basic phonemes, sperm whales utilize a finite acoustic inventory to produce an effectively unlimited expressive range.
The Discovery of Vowels: A phonetics analysis by UC Berkeley revealed that individual whale clicks contained precise frequency modulations that look structurally identical to human vowel sounds, such as the a in father or the e in see.
Convergent Evolution and a Theory of Mind
A Feature of Intelligence: Because humans and whales haven't shared a common ancestor for hundreds of millions of years, the appearance of vowels is a profound example of convergent evolution. It implies that vowel-based communication is not a human anomaly, but a fundamental characteristic of high intelligence.
Pragmatic Context & Relationships: Sperm whales have the largest brains on Earth and live in deep, matrilineal families. Their conversations last up to an hour and exhibit turn-taking, rhythm, and adjustments based on who they are addressing, requiring a baseline "theory of mind" previously thought unique to humans.
The Moral Implications of Translation
Structure vs. Meaning: While AI has mapped the complex architecture of whale grammar, researchers have not yet translated a single word. The next phase relies on cross-referencing acoustic data with specific behaviors to decode actual meaning.
Environmental Damage to Culture: If sperm whales possess a multi-generational spoken culture, underwater industrial noise pollution isn't just ecological damage, it represents the systematic disruption of an ancient society's communication space.
That’s a wrap!
We'd love to hear your thoughts on today's email!Your feedback helps us improve our content |
Not subscribed yet? Sign up here and send it to a colleague or friend!
See you in our next edition!
Gina 👩🏻💻