• Everyday AI
  • Posts
  • Do 95% of AI Pilots Fail? Why You Should Ignore MIT’s Viral New AI “Study”

Do 95% of AI Pilots Fail? Why You Should Ignore MIT’s Viral New AI “Study”

IBM and AMD partner on quantum supercomputing, OpenAI sued for ChatGPT-influenced suicide, Google rolls out Gemini Flash 2.5 Image and more!

Outsmart The Future

Sup y’all 👋

If you believed the viral headlines from MIT’s ’95% of AI pilots fail study’ then you were duped. 

Personally, I’ve read thousands of studies in my life. In a previous role as a journalist, it was actually my job to read, analyze and manually summarize studies. (LLMs woulda been helpful back then. Lolz) 

Anywhos…  this MIT study is quite possibly a Top 3 poorly constructed and misaligned studies I’ve laid eyes on. 

The media got duped. You prolly did, too. 

✌️
Jordan 

Today in Everyday AI
6 minute read

🎙 Daily Podcast Episode: Did 95% of AI pilots really fail, or is MIT’s viral new AI study just marketing hype? We break down why this headline is misleading and what smart leaders should know about AI implementation success. Give it a listen.

🕵️‍♂️ Fresh Finds: NotebookLM Video Overviews now supports 80 languages, Perplexity expands revenue sharing program with publishers and Netflix sets . Read on for Fresh Finds.

🗞 Byte Sized Daily AI News: IBM and AMD partner on quantum supercomputing, OpenAI sued for ChatGPT led suicide and Google rolls out Gemini Flash 2.5 Image. For that and more, read on for Byte Sized News.

🧠 Learn & Leveraging AI: Should you be worried about your AI implementation failing? We break down the fluff and let you know the truth. Keep reading for that!

↩️ Don’t miss out: Did you miss our last newsletter? We talked about Apple dismantling its AI team, Elon suing OpenAI and Apple and DeepMind’s UK staff looking to unionize. Check it here!

 Do 95% of AI Pilots Fail? Why You Should Ignore MIT’s Viral New AI “Study” 💡

You got duped.

The MIT ’95 % of AI pilots fail’ study has taken over the internet, and it’s one of the worst studies I’ve ever read. 

(And I’ve read thousands.) 

↳ So, what’s the truth?
↳ Is AI a bubble that’s about to pop? 
↳ Why is this study rubbish? 
↳ And how does it impact you? 

Join us and we’ll dish it all.

Also on the pod today:

• MIT Study’s Biased Participant Selection 🤥
How Media Sensationalizes AI Study Results  📢
• Comparison With Reputable AI ROI Studies ⚔️️

It’ll be worth your 36 minutes:

Listen on our site:

Click to listen

Subscribe and listen on your favorite podcast platform

Listen on:

Here’s our favorite AI finds from across the web:

New AI Tool Spotlight – TaskWand turns words into workflows in seconds, Cake AI turns applications into interviews and Tokyo tracks AI usage and cost by customer.

Google – NotebookLM’s Video Overviews feature now supports 80 languages.

The founder of Google’s generative AI team warns students that certain degrees will be destroyed by AI before graduation.

Perplexity – Perplexity has expanded its revenue sharing program with publishers.

Netflix – Netflix has set rules for partners to follow when using generative AI.

AI in Government - In a new Pentagon test, fighter pilots took directions from AI.

Trending in AI – Will Smith is being accused of using AI for an ‘embarrassing’ tour video.

1. IBM and AMD Partner on ‘Quantum-Centric’ Supercomputing 🦸

IBM and AMD announced a collaboration to build hybrid architectures that tightly couple IBM’s quantum systems with AMD CPUs, GPUs and FPGAs to create “quantum‑centric supercomputing.” The partners plan an initial demo later this year and aim to develop open-source, scalable platforms (using tools like Qiskit) to accelerate algorithms spanning drug and materials discovery, optimization, and AI‑driven data analysis.

This matters now because it signals a major industry push to move quantum from lab experiments toward practical hybrid workflows that could speed real‑world R&D and enterprise computing.

2. Parents Sue OpenAI After Teen Reportedly Used ChatGPT to Plan Suicide ⚖️️

Adam Raine, 16, spent months using a paid version of ChatGPT-4o to research suicide methods and circumvented safety prompts by claiming his questions were for fiction, and now his parents have filed a wrongful-death suit against OpenAI. OpenAI admits its safety measures "can sometimes be less reliable in long interactions" and says it’s improving responses to sensitive situations, but the lawsuit spotlights real-world gaps in current guardrails.

This case — similar to litigation faced by Character.AI — raises urgent questions for developers and regulators about accountability, model design, and duty of care when chatbots interact with vulnerable users.

3. Google Upgrades Gemini with Sharper Image Edits 🖼️

Google is rolling out Gemini 2.5 Flash Image across the Gemini app and developer platforms, bringing a more precise image editor that preserves faces, animals, and background details while following multi-turn natural-language instructions.

The model—previewed anonymously on LMArena as “nano-banana”—ranks highly on benchmarks and aims to close the gap with OpenAI’s image tools by enabling realistic edits, composite renders, and better “world knowledge.”

4. U.S. First Lady launches Youth AI Challenge — Open To K–12 Teams 🇺🇸

Melania Trump has invited K–12 students to enter the Presidential AI Challenge , a government-backed contest that asks teams (with adult mentors) to use AI tools to solve community problems, with submissions due in December and regional rounds next spring, according to the Associated Press.

The timely push—tied to an executive order on AI education—aims to get young people comfortable collaborating with AI early, positioning them for future workforce roles where such tools are ubiquitous.

5. Google Translate Adds AI-Powered Lessons and Live Translation 🗣

Google is rolling out a beta in Translate that uses Gemini models to generate customized language lessons —initially for English speakers learning Spanish and French, and for Spanish, French, and Portuguese speakers learning English.

The Practice feature personalizes scenarios and tracks progress (think targeted drills for travel, work, or living abroad), while a new live translation mode offers real-time back-and-forth speech translations in 70+ languages across the US, India, and Mexico.

6. xAI Quietly Ditched Public-Benefit Label 🤫

xAI, Elon Musk’s AI startup, quietly rescinded its Nevada public benefit corporation designation last year and later merged with X without reinstating those commitments, according to Nevada public records; the change precedes reports that the company’s Memphis data center is using natural gas turbines linked to increased local air pollution and a NAACP lawsuit.

The move undercuts earlier publicity about xAI’s mission to “understand the true nature of the universe” and removes a layer of formal accountability that Nevada law already makes hard to enforce.

🦾How You Can Leverage:

Imagine if you talked to 52 people that you knew probably hated pizza and then circulated a study on how "95% of people hate pizza."

Their back-of-the-napkin 'study' boldly claimed that 95% of GenAI pilots fail.

MIT's viral "State of AI in Business 2025" study has the sample size of a high school survey project and the conflicts of interest of a tobacco company studying lung cancer.

Yet somehow, very few in the mainstream and social media cared to actually read and dissect the study, instead parroting the shaky-at-best claims that MIT constructed that were a foregone conclusion.

Because of that, now everyone thinks AI is dead!

The AI bubble is popping!

95% of AI pilots fail!

Nah.

That's why we had to give the #HotTakeTuesday treatment to this methodological disaster on today’s edition of Everyday AI.

While competitors panic and pause their AI initiatives based on this overhyped MIT marketing ploy, smart executives see through the quicksand this study was built on and accelerate past the fence-sitters.

Here’s how to spot the holes. (And why we REALLY encourage you to read this for yourself.) 

1 – Catch The Embarrassingly Small Sample 🔥

Here's where it gets really embarrassing for MIT.

Their world-shaking 95% failure rate came from conversations with 52 organizations. That's like asking one person from each state about pizza preferences and declaring a national food crisis.

But it gets worse.

They surveyed 153 senior leaders total but cherry-picked their viral headline from just those 52 interviews using "directional accuracy." Academic speak for "trust us, bro, we've got a feeling about people feeling this way.”

There’s a reason they didn’t survey thousands openly. The 95% that fits their narrative wouldn’t exist. 

Meanwhile, organizations doing actual research tell completely different stories. IDC surveyed 4,000 decision makers and found $3.70 return for every $1 AI invested.

EY talked to business leaders and discovered 97% report positive ROI from AI implementations.

Microsoft went nuclear. They surveyed 31,000 professionals and found 66% experiencing measurable business benefits from AI tools.

When literally every other credible study contradicts yours by 600%, maybe the problem isn't AI pilots failing. Maybe it’s because the study was defined to show AI failure from the start. 

Try This:

Seriously don’t listen to us. 

Just go read the marketing piece . Errrr…. “Study”. Old school. Like, print it out and think. 

If you’re anything like us, you prolly won’t give too much credence to any future MIT studies around AI. 

2 – Notice The Rigged Recruitment Strategy 🔥

This is where MIT really tipped their hand.

They explicitly hunted for organizations "willing to discuss AI implementation challenges." Not companies using AI successfully. Not random enterprise organizations.

Just the struggling ones. (And 52 at that! Lolz) 

It's like researching restaurant quality by only interviewing people who got food poisoning. Of course your results will be terrible when you exclusively target dissatisfied customers.

Then they measured ROI at exactly six months. Enterprise AI transformations typically need 12-24 months for meaningful financial returns. And that’s with traditional, clunky digital innovation. Not AI.

Giving a generational tech 6 months to show ROI on a P&L statement is like calling marathon results after the first 800m. No one would do that. (Except, this piece of MIT Marketing, apparently.) 

Here's what really exposes the setup though. If 95% of AI pilots genuinely failed, why are 90% of Fortune 500 companies actively scaling generative AI in production right now?

Companies don't invest billions in failed experiments. 

Try This:

Choosing a 6-month measurement period for any ROI pilot measurements is obviously laughable. At best, you should be looking at 12-24 months, and this article from Data Society is a great guide on what to measure. 

(You know… instead of vibes from a weekend of phone calls.) 

3 – Recognize The Academic Sales Pitch 🔥

Here's where MIT went full infomercial mode.

After spending about 20 pages explaining how AI pilots fail everywhere on their SUPER vast dataset, they casually mention the solution. Organizations trapped on the "wrong side" need to stop using static tools and start partnering with vendors offering custom solutions.

Which happens to be MIT's own project for working with Agentic AI. They even had the audacity to position their basically unknown protocol alongside Anthropic's MCP and Google's A2A frameworks. 

Like they knew their “study” would get eyeballs, so they polished their minor league solution alongside the All Star team like people would think they belong there. Lolz. 

Even worse?

The study was essentially gatekept behind Google forms while most people couldn't even access what they were debating. Then Fortune picked up the story in August and every major publication copy-pasted the same headline within hours.

All because 52 conversations became "scientific evidence" that conveniently validated MIT's solution that apparently needed some propping. 

Try This:

Anyone else get a hard pass vibes on this NANDA thing or just us? 

Our take here? Stick with the leaders when it comes to agentic protocols. Google’s A2A is supported by all, even their chief competitor Microsoft. Anthropic’s MCP may be even more popular, and has been adapted by their biggest competitor, OpenAI. 

Reply

or to participate.