The iron horse rides again
We who work in marketing know how important names and terminology are for branding and selling. Sometimes an industry rallies around a label for an entire category. An “EV” is the established acronym for an electric vehicle and nobody tries to call them “gasoline-free” or “greencars”. Earlier this year, the United States FDA issued guidelines on commercial applications of the word “milk” for plant-based drinks. And the term “AI” is universally used by tech leaders, scientists, and commentators when referring to machine learning platforms and products.
But does "artificial intelligence" accurately represent LLMs such as ChatGPT and diffusion algorithms such as Dall-E? Have we been parroting the term “AI” without analyzing its meaning and the consequences of using it?
History provides an enlightening example from a different ground-breaking technology. When the steam locomotive was invented in the 1820s, it was a marvel. People felt the need to call it something they could relate to, maybe something less frightening, so it was quickly dubbed the iron horse.
In hindsight, we can smile at the quaintness of the term “iron horse”. But nobody in 1840 thought that a steam engine was the same thing as a galloping animal. The problem is that today a whole lot of people have been fooled into believing that AI is the same thing as a thinking mind. Terms such as “neural network” only enhance this illusion.
We can nitpick about the definition of intelligence (even expert scientists don’t agree on it) and LLMs are very good at producing convincing text (the “chat” in ChatGPT), but an LLM doesn’t have neurons or an internal model of the world any more than a steam engine has lungs or a beating heart. These platforms have passed the Turing test while simultaneously proving that humans are so bad at distinguishing fake minds from real ones that the Turing test has become pointless.
Several systems have been developed that perform specific tasks far better than any human, such as playing chess. But there’s a chasm between a machine performing a task super efficiently, such as locomotion, calculation, or playing a game with clear rules, and it being intelligent like a person.
The word “intelligence” in the term “artificial intelligence” is as ridiculous as the word “horse” was in the term “iron horse”. Luckily, an alternative term to “AI” already exists: machine learning (ML). We really should start using it.
You might be saying to yourself: Who cares whether LLMs have human-style intelligence? Shared terminology and overarching narratives matter when thinking about safety, law, policy, and regulation. And yet we have all (myself included) blindly adopted the narrative and terminology of the tech industry. Companies with products to sell, like OpenAI, X, Google, and Meta, profit from this “intelligence illusion” in two ways: a) they can charge more for access because the products are over-hyped, and b) they can influence government regulation to capture the market from newcomers by stoking unfounded fears.
We shouldn’t be tempted to conclude that AI is benign because it lacks genuine intelligence. But let’s stop fooling and scaring ourselves into believing that machine learning systems are in any way sentient. If AI is a threat to society, we need to understand our enemy properly.
My prediction? By the end of this century the term "artificial intelligence" will seem as quaint as "iron horse" does now. So subscribe to Discomfort Zone today to be sure that you’re still receiving it in 2099!
Pinterest searches get more inclusive
Yesterday Pinterest launched a new feature that allows users to filter image searches by body type, in an effort to promote inclusivity on its platform in a week when litigation against Meta by US attorneys general from 33 states accuses the social media giant of routinely collecting data on users under 13, as part of a wider lawsuit about harms caused to children’s mental health.
Pinterest’s new tech is powered by (you guessed it!) machine learning, and comes on the heels of a previous rollout of a skin tone range filter in the US, Canada, Britain, Ireland, Australia, and New Zealand that the company claims resulted in a 70% annual increase in users saving Pins from across the skin tone ranges in their feeds. Pinterest’s Megan D’Alessio, manager of inclusion and diversity, commented that body dissatisfaction is prevalent for young people, but the issue is far worse for women as “50% of adolescent girls are unhappy with their bodies compared to 31% of boys.”
Music in the Key of AI Major
Alphabet’s YouTube has partnered with Alphabet’s DeepMind machine learning model to create two new ways to generate music. Lyria features a tool called Dream Track that allows users to produce 30-second-long vocal tracks from text prompts. Each 30-second “song” will feature the AI-generated voice and musical style of artists such as Alec Benjamin, Charlie Puth, Charli XCX, Demi Lovato, John Legend, Sia, T-Pain, Troye Sivan, and Papoose.
The second system, called Music AI Tools, enables users to input audio using their own voice (imagine humming a short melody) and then transform this rough track into a polished version played on an instrument of their choice, to which they can add rhythm tracks, full instrumentation, and production effects. Lyria is in beta mode for a small test group, but Google promises participating artists that these tracks cannot be copied without tracing the source through DeepMind’s revolutionary SynthID audio watermarking system. Check out the demo video here.
Greener Skies on the Horizon?
Sir Richard Branson was one of the few passengers on board a Virgin Atlantic 787 from London to New York this week that was billed as the world’s first fossil-fuel-free transatlantic jetliner flight. Both engines were powered by sustainable aviation fuel (SAF) derived from waste fats like used cooking oil and animal fat, mixed with synthetic aromatic kerosene made from dextrose extracted from corn during the production of animal feed. Branson seems to have missed the opportunity to launch an Extra Virgin Atlantic olive oil brand.
Sounds great, right? Well, sustainable aviation fuel represents around 0.1% of aviation fuel globally and it will be almost impossible to produce enough to meet sustainability targets set by the US and UK governments. And that isn’t the only cloud on the horizon. It was revealed this week that Saudi Arabia has a covert program to expand the use of fossil fuel-powered planes in Africa and elsewhere to maintain demand as rich countries switch to clean energy. The Saudi Oil Demand Sustainability Program (ODSP) aims to subsidize the adoption of oil in a wide range of sectors across the continent. This will, of course, also fuel the climate catastrophe. The article quotes Mohamed Adow, director of the Power Shift Africa think tank, who said: “The Saudi government is like a drug dealer trying to get Africa hooked on its harmful product.” So it seems like using cooking fat for jets won’t stop us all from being fried someday soon.
Sarah Lolley, medical writer
Q. As an expert medical writer who specializes in medical education and has partnered with some of the world’s leading pharmaceutical companies, how do you see your role at a time when many people are skeptical about what they hear from pharma brands and public health agencies?
A. I can understand a certain amount of skepticism; I think it comes from a lack of appreciation of just how highly regulated pharmaceutical research is and how specialized the people who work in this field are.
I'm also surprised when I hear criticism directed at the companies whose mission it is to advance medical science but not at those whose products can be harmful to our physical or mental health.
My advice to those who feel vitriol for the pharmaceutical industry is to aspire never to need any of their products. We can stave off a lot of chronic diseases by being sober and smoke-free, eating a low-salt, low-fat, plant-rich diet, and getting vigorous exercise every day.
But if, despite a healthy lifestyle, some illness should befall them – if cancer should grow in their bodies or if their genes should turn against them or if they should catch a virus that their immune systems are not strong enough to fight alone – they can turn to the pharmaceutical industry for treatments that been studied extensively and are known to be effective.
Sarah is a medical writer specializing in CME, with over twenty years of industry experience internationally. She is passionate about science communication, storytelling, and innovation in education. Sarah is also fascinated by puzzles and how we solve them.
Confident Asymmetry
A certain degree of discomfort is a common feeling to have when we’re faced with a person who looks significantly different from the norms of society and biology. (This discomfort doesn’t mean that bigotry and discrimination are okay!) We can all recognize our reflexive reactions if we make an effort, and it doesn’t take much psychological digging to realize that our own fears have provoked a negative emotional response. This is why it’s important to have representation in the media of people whose appearance doesn’t align with the majority in a community. As covered in the “FOMO food” item about Pinterest above, it’s also important that people should feel comfortable in their bodies, whatever their shape.
This article tells the story of how discomfort was dealt with by award-winning British architect Katy Marks, who had a single mastectomy following a breast cancer diagnosis. She has just launched the –unobra line of asymmetrical lingerie and swimwear for women who have one breast. Marks sourced sustainable materials for unobra and has worked hard to maintain affordability. The article quotes her as saying, “My experience is that many people don't even notice, but when they do, they accept it, and why shouldn't they? Why should we feel ashamed or try to hide our bodies after we have had cancer, as though the only way to be feminine is to be a 'normal woman'? We need to shift into the 21st century and recognize that the diversity of our bodies is the norm. The more we see it, the more we can be it.”
Obama talks AI
Are you overwhelmed by all the tech gurus, businesspeople, and journalists talking about AI these days? I get it. So when I heard that Barack Obama had been interviewed on the subject by The Verge’s Nilay Patel for the Decoder podcast I decided it might be worth a listen. It was. Obama has the time, access, and (genuine) intelligence to explore the subject deeply and share his opinions eloquently.
The former president also discusses free speech and the future of the internet. If you prefer to watch two men talk into large microphones, the YouTube version is here.
Hands Up If You Still Love Video Meetings
For those of you who didn’t get all zoomed-out during the pandemic, Google has just launched the ability to raise your hand during a Meet session instead of clicking the little hand icon. This feature might not live long and prosper. What do you think?
Send me your comments on Discomfort Zone at john@johnbdutton.com and please connect with me on LinkedIn if you haven’t already.