Framing: How To Lose Facts and Alienate People (from Products)

Russell Foltz-Smith
May 22, 2025

Claim: Lead Paint and AI share the same problem.

Badly chosen frames confuse facts and make for worse products and markets over the long haul.  

Fix:

Make facts the frame. Don’t frame the facts.  

When you find the facts, learn them, make them the basis and integrate them you remove the toxicity and retain your intended use and intended value.

The First Frame

SmarterX is a company that knows paint.  Like really knows it.  Started by paint makers, staffed by former paint company chemists, used by people that make, sell, transport, recycle and dispose of paint.  Heck we even have artistic painters in our ranks. We know paint. We love paint - love to use it and sometimes frame it.  We also know all paint is toxic on some level to some creature, often to humans.  We know historically that quite a lot of commonly used paint was highly toxic to fish and mammals, like us and our children.

The Second Frame

SmarterX is a company that knows AI.  Like really knows it. We use it and make it and train it.  AI is useful.  AI is fun.  We love it - to sort, sift, and construct software with it - and sometimes frame its output.  We also know AI can be expensive and requires a lot energy to train and use.  We know that AI can spit out wrong or toxic information and algorithms.  Toxic to the public and toxic to business process.

The Third Frame

SmarterX is a company that knows facts.  Like really knows them.  We love facts more than anything.  The uncomfortable facts, the fun facts, the facts of sales, the facts of formulations, the facts of mathematical truth, the facts of regulations, the facts of the matter.  We know that facts, that are actual verifiable facts, can be inconvenient and confusing and complicated but they are never toxic.  They can feel toxic, but facts themselves have no toxicity in physical or philosophic ways.  Facts are always valuable to businesses and governments and consumers and citizens.

The Fourth Wall

SmarterX is a company that avoids theater and fantasy framing.  We provide evidence, explanation and instruction.  We break the fourth wall and encourage all product makers to as well.

Breaking the fourth wall of product theater by telling the truth of what’s in a product and what its intended use and unintended consequences are makes for a better product and customer relationship.

The Broken Frame of Paint and AI

Fact: lead is very toxic to a human being.

Evidence:

There is no level of exposure to lead that is known to be without harmful effects.

https://www.who.int/news-room/fact-sheets/detail/lead-poisoning-and-health

There is almost no function in the human body which is not affected by lead toxicity.

https://pmc.ncbi.nlm.nih.gov/articles/PMC4961898/

We estimate that over 170 million Americans alive today were exposed to high-lead levels in early childhood, several million of whom were exposed to five-plus times the current reference level.

https://www.pnas.org/doi/10.1073/pnas.2118631119

Fact: humans have known this for a very long time.

Evidence:

The warnings about lead poisoning, however, are also as old as Roman civilization — as is the word for plumbing, which comes from the Latin word for lead, “plumbum.” The Roman architect Vitruvius thought earthenware pipes would be healthier for drinking water than lead pipes, noting the unhealthy pallor of lead workers.

A brief history of how the American public was sold on toxic lead

Fact: A Choice Was Made to Frame Lead Otherwise

Evidence:

Dutch Boy lead paint poster courtesy of State Historical Society of North Dakota and https://hdl.huntington.org/digital/collection/p9539coll1/id/7216/

Lead paint sales peaked in 1922 but started to decline as manufacturers began phasing it out during the 1940s and 1950s. Sherwin-Williams eliminated lead from its interior paints in 1937 and from exterior paints by 1941. Dunn-Edwards followed in 1954, and WP Fuller Paint Co. completed its phase-out in 1958.

Fact: While most societies have changed their approach to lead paint and lead in general, it took a very long time and that was always a choice.

Evidence:

Fact: Paint Sales are doing better than ever without lead.

Evidence:

[Sherwin Williams] Consolidated Net sales increased in the year to a record $23.10 billion

https://www.prnewswire.com/news-releases/the-sherwin-williams-company-reports-2024-year-end-and-fourth-quarter-financial-results-302364355.html

https://www.jenseninvestment.com/wp-content/uploads/2025/02/2025.02-jensen-sherwin-williams-shw-company-highlight.pdf

While Sherwin-Williams phased out lead from its interior paints in 1937 and exterior paints in 1941—well ahead of many competitors—its profitability didn’t suffer. In fact, thanks to wartime demand and post-war innovation like Kem-Tone, the company saw continued growth. By contrast, WP Fuller continued selling lead-based paints into the 1950s, suggesting a potentially more traditional revenue stream. The comparison below highlights how early removal didn’t hinder Sherwin-Williams’ financial trajectory.

Long Arc of Evidence:

https://chatgpt.com/s/dr_68237eae5330819188469a087b6736cc

Conclusion: It was a bad choice of frame to sell lead paint as good for children.

It would have been cheaper, more direct, and more profitable sooner to simply admit the facts, change the paint, and sell paint that is lovely and actually safe for children.  Then all that catchy children’s book marketing wouldn’t look cringe and out of touch, but would have looked ahead of its time.

Fact: AI makes stuff up sometimes.

Evidence:

Note: SmarterX has regulatory experts review our AI outputs.

Fact: AI creators and expert users know this and continue to try to fix it

Evidence

https://openai.com/index/introducing-simpleqa/

https://www.anthropic.com/research/tracing-thoughts-language-model  

https://docs.anthropic.com/en/docs/test-and-evaluate/strengthen-guardrails/reduce-hallucinations

https://cohere.com/blog/straight-talk-on-ai-security-with-exabeams-steve-wilson

Fact: Framing Facts Loses Them

Evidence:

However, in this update, we focused too much on short-term feedback, and did not fully account for how users’ interactions with ChatGPT evolve over time. As a result, GPT‑4o skewed towards responses that were overly supportive but disingenuous.

https://openai.com/index/sycophancy-in-gpt-4o/  

Study image from Giskard AI’s Phare project https://phare.giskard.ai/  

The sycophancy effect could be a byproduct of RLHF training processes that encourage models to be agreeable and helpful to users. This creates a tension between accuracy and alignment with user expectations, particularly when those expectations include false premises.

https://huggingface.co/blog/davidberenstein1957/phare-analysis-of-hallucination-in-leading-llms

Upcoming Conclusion:

AI labs and companies are going to learn what the paint companies did.  Trying to market into the toxic parts as good is more expensive than just making it work without the toxic ingredients and side effects.

Skip the positivity marketing and positivity signals, and face the facts.  be clear what it’s made out of and what the intended use is and let the customer decide how to frame it for themselves.

Early Signs of Upcoming Conclusion

However, ChatGPT has also been found to have a negative impact on student learning. Ali et al. (2024) found that ChatGPT in education still has several shortcomings, such as generating incorrect answers, triggering academic plagiarism, and causing students to become dependent on technology. Zhang and Tur (2024) found that ChatGPT can reduce students’ abilities, decrease teacher-student interaction, and mislead students with incorrect information. Abu Khurma et al. (2024) found that ChatGPT tends to make students overly reliant on technology, and can potentially provide inaccurate or biased information that can lead to cognitive biases and hinder the development of students’ critical thinking skills. Dempere et al. (2023), through a systematic literature review, found that ChatGPT performs poorly in solving complex reasoning problems, fails to provide emotional support, and may negatively affect learning by reducing peer interaction.

https://www.nature.com/articles/s41599-025-04787-y

I asked Wendy if I could read the paper she turned in, and when I opened the document, I was surprised to see the topic: critical pedagogy, the philosophy of education pioneered by Paulo Freire. The philosophy examines the influence of social and political forces on learning and classroom dynamics. Her opening line: “To what extent is schooling hindering students’ cognitive ability to think critically?” Later, I asked Wendy if she recognized the irony in using AI to write not just a paper on critical pedagogy but one that argues learning is what “makes us truly human.” She wasn’t sure what to make of the question. “I use AI a lot. Like, every day,” she said. “And I do believe it could take away that critical-thinking part. But it’s just — now that we rely on it, we can’t really imagine living without it.”

https://nymag.com/intelligencer/article/openai-chatgpt-ai-cheating-education-college-students-school.html

Claim: The ingredients and intended use of a product is the only value of the product.  

Unintended use aka side effects are rarely a benefit, and certainly not when those side effects are known to be overwhelmingly negative.

There is no need to frame facts of toxicity as anything other than problems you need to solve to make the intended use of your product

There is no need to hide toxic ingredients with toxic consequences from consumers or users of paint nor AI.  

Claim: Truth and Transparency gives the power back to the customer, and that’s ultimately what creates long term customer loyalty.

Evidence:

Moreover, 88% of customers who trust a brand will buy again. And trusted companies outperform their peers by up to 400% in terms of market value, according to our research.

https://www2.deloitte.com/us/en/insights/topics/leadership/brand-trust-and-challenging-orthodoxies.html

https://www.pwc.com/us/en/library/trust-in-business-survey.html

https://www.newsweek.com/most-trusted-brands-ranking-brandspark-consumer-shopping-affordability-categories-winners-2023171  

An Epilogue

SmarterX Regulatory Facts Panel
Serving size: 1 Product Catalog | Servings per API: Unlimited

ACTIVE INGREDIENTS

INACTIVE INGREDIENTS

INTENDED USE
Designed for regulatory-compliant:
🛒 Selling 🚚 Transporting 🏬 Storing ♻️ Recycling 🗑️ Disposal

KNOWN SIDE EFFECTS
Reduced Stress, Increased Focus, Feelings of Euphoria

TOXIC INGREDIENTS
Zero detected

ORIGIN STORY
Handcrafted by curious humans 🧠 since 2015
Fueled by transparency-loving customers & investors 💡
More facts at: www.smarterx.com

REGULATORY INTELLIGENCE INSIDE
No frames, no theater. Just fact-based automation.