Media Style Guide
How to talk about "AI".
Inspired by the Vision Zero Vancouver Style Guide, we present some of the ways we think are more helpful for parents, guardians, educators and students to talk about "AI".
This is a real product. In this case "AI" means it has a sensor that checks if you missed a spot.
What do you mean by "AI"?
"AI" is a term that is being applied to everything. Robot vacuum cleaners now have "AI". Upscaling, a feature that has been around for a decade, is now "AI Upscaling"
that upscale images are now AI TV" Techniques and product features that have existed for decades are being rebranded as "AI".
This vagueness helps companies. Say for example, scientists create a research tool, trained on a large number of MRI scans of cancers, and use it to identify potential cancer in new images. If that is labelled as "AI", then when a company creates a chatbot and calls it "AI", it is associated with that same level of cutting-edge technology and usefulness.
If people complain about "AI" in the form of chatbots in classrooms, proponents of the product will say "AI is used in research to identify cancer, we must make sure students are using it too."
This makes as much sense as saying that because radiation is used to treat cancer, students should be playing with uranium in classrooms.
- ❌ "AI is now in classrooms."
- ✅ "Chatbot products like Copilot are now in classrooms."
- ❌ "AI can identify cancer."
- ✅ "Researchers used a program to identify cancer."
- ✅ "Researchers used an algorithm to identify cancer."
- ❌ "AI can now make movies."
- ✅ "People can use products to create videos."
Question premises
Question that something that uses "AI" is inherantly better than something that does not.
Questions to ask
LLM products's perceived benefits are often unquestioned. If a school or government body is proposing to make AI available to students, ask how those students would benefit from this plan. ask what how students benefit by ai being put in classrooms
The first question when talking about AI in classrooms is How will this product help students?
That can be approached from many directions:
- What studies have been done using this product and its effects
The human aspect of this is also important to cover.
- Have you talked to parents and students about the story?
ask who
- Where is the information stored?
- What is the data retention policy?
- Are questions used to train?
- What is the copyright status of anything created by using the product?
"Inevitable"
Don't anthropomorphise
Microsoft and OpenAI love to talk about their products as if they are a person. Almost all chatbots speak using first-person pronouns, saying "I did this."
It is hard to talk about a product that carries out tasks without acccidentally speaking about it as if it has agency.
When people use language that treats LLM products like a person, it makes the product seem more capable, and degrades the value of actual humans.
Centre the human, not the product
- ❌ "AI can solve this problem."
- ✅ "People can use an LLM to an answer."
LLMs cannot solve problems. They can output text which may be the same as a correct answer.
Just like rolling dice might come up with the answer to 3+3.
Understanding
- ❌ "Copilot understands what you're saying."
- ✅ "The AI output text."
- ❌ "ChatGPT thought..."
- ✅ "ChatGPT"
LLMs are not capable of understanding, no more than the autocomplete on your phone can understand the content of your email.
Hallucination
- ❌ "One problem with AIs is their frequent hallucination"
- ✅ "One problem AI up
Humans hallunciate, LLMs create output when given an input. Sometimes that output matches reality, sometimes it does not. By using a complex term like "hallucinate", it hides the simple truth that LLMs are designed to make things up.