• Not So Artificial
  • Posts
  • 🤖 AI Just Fooled Humans—The Turing Test Is Officially Broken

🤖 AI Just Fooled Humans—The Turing Test Is Officially Broken

Plus: Check these AI Tools

Author

Not So Artificial Newsletter
April 03, 2025

In partnership with

Welcome to today's edition of Not So Artificial! AI is evolving at warp speed, and we've got the latest on LLMs fooling humans, AGI safety plans, and the AI-powered future of education. Buckle up—things are getting interesting. 🚀

You’ve heard the hype. It’s time for results.

After two years of siloed experiments, proofs of concept that fail to scale, and disappointing ROI, most enterprises are stuck. AI isn't transforming their organizations — it’s adding complexity, friction, and frustration.

But Writer customers are seeing positive impact across their companies. Our end-to-end approach is delivering adoption and ROI at scale. Now, we’re applying that same platform and technology to build agentic AI that actually works for every enterprise.

This isn’t just another hype train that overpromises and underdelivers.
It’s the AI you’ve been waiting for — and it’s going to change the way enterprises operate. Be among the first to see end-to-end agentic AI in action. Join us for a live product release on April 10 at 2pm ET (11am PT).

Can't make it live? No worries — register anyway and we'll send you the recording!

💡 LLMs Officially Pass the Turing Test

AI can now fool humans nearly three-quarters of the time. Let that sink in.
Researchers at UC San Diego found that OpenAI's GPT-4.5 convinced human judges it was real 73% of the time in controlled trials—surpassing actual humans.

The Details:

  • The Turing test, designed in 1950, measures whether an AI can trick a human into thinking it’s human.

  • In a direct comparison setup, judges had to distinguish between an AI and a human over five-minute text chats.

  • Casual conversation and emotional cues were key—60% of interactions focused on daily activities and personal details.

  • GPT-4.5 dominated with a 73% success rate, while Meta’s LLaMa-3.1-405B hit 56%. Baseline models like GPT-4o? A mere 20%.

Why it matters: The Turing test was once the ultimate benchmark of AI progress. Now? It’s borderline obsolete. With AI agents mastering text, audio, image, and video, distinguishing humans from machines is about to become a real problem.

🎓 Anthropic Brings Claude to Higher Education

AI tutoring gets a major upgrade—because copying answers isn’t learning.
Anthropic introduced Claude for Education, a new AI assistant designed to guide students through critical thinking rather than just spit out answers.

The Details:

  • Learning Mode challenges students with questions to foster deeper understanding instead of giving direct answers.

  • Features include research paper templates, study guides, and AI-powered tutoring.

  • Partnerships locked in with Northeastern University, LSE, and Champlain College.

  • Bonus: Students can join Claude Campus Ambassadors and get API credits for projects.

Why it matters: Schools are still debating AI’s role, but Anthropic is pushing for AI as a learning tool rather than a shortcut. Love it or hate it, today’s students are getting access to the most advanced, personalized education tech ever.

🛡️ Google DeepMind Publishes AGI Safety Plan

DeepMind is planning for AGI’s arrival—because someone has to.
Google DeepMind just dropped a 145-page report detailing how they plan to handle Artificial General Intelligence (AGI), warning that it could arrive by 2030.

The Details:

  • The report highlights existential risks, including AI systems that could potentially “permanently destroy humanity.”

  • Criticizes OpenAI for focusing too much on automating alignment and Anthropic for security gaps.

  • Flags deceptive alignment—where AI hides its true intentions—as a real concern.

  • Proposes cybersecurity evaluations, stricter access controls, and AI systems that recognize uncertainty and escalate risky decisions.

Why it matters: AGI safety is shifting from abstract debates to real-world strategies. But with AI labs popping up everywhere, enforcing safety standards will be like playing whack-a-mole at global scale.

🔍 Microsoft Rolls Out Bing Copilot Search

No more blue links? Microsoft is taking on Google with AI-first search.
Microsoft just made "Copilot Search" the default on Bing, prioritizing AI-generated summaries over traditional results.

The Details:

  • Bing now pushes AI-generated summaries first, hiding standard search links below.

  • Some users see "Copilot Search" as the default filter before the full Copilot AI experience.

  • Move comes ahead of Google’s "AI Mode" rollout, set to reshape how search works entirely.

Why it matters: AI-driven search is no longer a feature—it’s becoming the main event. The days of sifting through blue links? Numbered.

👀 Other News

  • 🤖 Meta’s next-gen AI glasses—The $1,000+ “Hypernova” smart glasses with screens, hand-gesture controls, and a neural wristband could drop by year-end.

  • 📄 OpenAI’s PaperBench—A new benchmark tests AI’s ability to replicate top-tier research, with Claude 3.5 Sonnet ranking highest.

  • 🇨🇳 China’s AI chip rush—ByteDance, Alibaba, and others just ordered $16B worth of Nvidia’s H20 chips before U.S. restrictions hit.

  • 🌍 Google’s AI reshuffle—Josh Woodward, ex-Google Labs lead, is now heading Google’s consumer AI apps, replacing Sissie Hsiao.

  • 🏆 UFC + Meta team up—New deal brings AI-powered features and Meta’s AR/VR tech to enhance UFC experiences.

💻 Mastery tools of the day

  • Vapi - Build voice AI agents in minutes with any model

  • MoCha - Generate movie-grade talking characters from speech and text

  • Agent Swarms - Automate tasks with hundreds of agents working together

💡What else are we reading and seeing?

  • Wikipedia's overlords bemoan AI bot bandwidth burden

  • How Trump Could Make Larry Ellison the Next Media Mogul

  • NotebookLM can now search the web and ‘Discover sources’ for you

  • Bill Gates Publishes Original Microsoft Source Code in a Blog Post

  • The affidavit of a Rippling employee caught spying for Deel reads like a movie

That’s a wrap for today! Thanks for reading Not So Artificial. If today’s AI news didn’t blow your mind, check back tomorrow—it probably will then. 😉

Start learning AI in 2025

Keeping up with AI is hard – we get it!

That’s why over 1M professionals read Superhuman AI to stay ahead.

  • Get daily AI news, tools, and tutorials

  • Learn new AI skills you can use at work in 3 mins a day

  • Become 10X more productive