My colleague Dan Bradshaw wrote a piece recently on AI transparency, something we’re committed to at Valliance. I wanted to share a quick example of how our own process and commitment to transparency really helps build trust, and how not using it well can erode trust quickly not only for each other, but future generations.
In a recent article I wrote on Shadow AI Governance, I originally included three examples of ‘shadow’ innovations instead of two. This was from about 20 I got Perplexity and Claude to research. The third was generated by Claude as; Text Messaging: The €100 Billion "Network Test" The Shadow Story: In the early 1990s, SMS was just a technical testing channel for Nokia and Ericsson engineers to check if mobile networks were working. Teenagers in Finland discovered they could send messages for 10p instead of making 50p voice calls. Network operators were horrified - kids were clogging up their diagnostic channels with "miss u" and "c u l8r".
I asked for two citations for each of my examples, and it told me it got this one from Wired and Wikipedia. However, I didn’t read the cited articles. I then polished the example up to this having verified some numbers differently and added my own thinking and words; Text Messaging: From system misuse to $80bn of revenue
SMS was originally deployed as a network testing tool for engineers. Finnish teenagers discovered they could send messages at a fraction of the cost of a voice call, and network operators tried to stop this "misuse" of their diagnostic system. But cheap, silent, private communication that parents couldn't overhear and allowed you time to think before responding, was exactly what people wanted. Today 23 billion texts are sent daily. It then went into our content review process, and the sources were re-checked, and they didn’t stack up. Perplexity (usually much better at finding stuff on the internet) told me that customer to customer messaging wasn’t the original idea, but that it found no evidence that initial use of it was by teens hacking the system or that it was originally solely for network testing.
So I went back to Claude and asked the same question. Here’s their reply;
Let me search for credible sources about the SMS origin story.
—Shows search terms and sources examined—
You're right to be sceptical. After searching for reputable sources, I cannot find solid evidence for the specific narrative that:
SMS was originally just a technical testing channel
Finnish teenagers discovered they could use it for cheap messaging
Network operators were horrified by teenagers "clogging up" diagnostic channels
What the sources actually show:
The Real SMS Story:
SMS was developed as part of GSM standards by the European Telecommunications Standards Institute in the 1980s, originally intended for customers to receive alerts from their carrier/operator Wikipedia
The first SMS was sent on 3 December 1992 by Neil Papworth, a test engineer, who sent "Merry Christmas" to a colleague Wikipedia
By 1996, the younger generation discovered SMS and it became part of youth culture, primarily due to lower cost compared to voice calls Xoxzo
The Finnish Connection:
Finland did have extremely high SMS adoption rates in the late 1990s
Matti Makkonen, a Finnish systems designer, is credited as "the father of SMS" for developing the concept in a Copenhagen pizzeria Finland Abroad
By 2000, 85% of Finnish households had at least one mobile subscription, with 60% of 9-12 year-olds and 90% of 13-16 year-olds owning mobile phones Injuve
The story appears to be a simplified narrative that conflates several facts: Finland's early SMS adoption, young people driving usage due to cost, and SMS using signalling channels. But the "teenagers hijacking diagnostic channels" angle seems to be an embellishment that's become part of the folklore rather than documented history.
I should revise that example to be more accurate to the historical record.
Yes Claude, exactly right! Had I let my misinformation out into the world, carefully crafted by me in my own voice, and not appearing to have any origin in AI, it could almost certainly have become part of the next generation of training data, and suddenly this conflated story would have become fact.
AI Transparency isn’t only about being honest with clients.
It’s also a responsibility we have to each other and future generations to use this stuff well, and not lose our actual history. If we lose trust in AI, or AI-enabled strategy or planning, then we’ll never realise its true value.
