- The AI Exchange
- Posts
- AI content worries got you down?
AI content worries got you down?
A safe content strategy to avoid Google down-ranking + Google Sheets GPT Use Case
β¨Quick reminder!β¨ Add [email protected] as a contact in your email settings to ensure these emails go straight to your inbox in the future!
Welcome to another edition of what weβre determined to make the best damn newsletter in AI. Here weβll break down AI topics that matter, open your mind to use cases, and keep you ahead of the curve.
Our #1 goal is to be useful. So please shoot us an email π© if you have questions or feedback, and especially if you implement something we share!
Here's what we're covering today:
Mitigating Google's potential down-ranking of AI content with this "safe" content strategy
Using GPT-3 directly in Google Sheets to clean messy data
Keeping you up to date on AI news
... and if someone forwarded this email to you, thank them π, and subscribe here!
Letβs get to it! π
TODAY'S PERSPECTIVE
Down-ranking risk shouldn't preclude you from using AI to generate content
Worried about AI detectors down-ranking or banning AI generated content?
Here's our perspective on what a "safe" AI content strategy looks like to avoid down-ranking risk (backed by technical understanding - like we always do).
1) Assume that unedited AI content will be detectable
AI models have already proven their ability to detect AI. So we should assume Google has the same. Solution - make sure you have human editors in any AI content creation process. (aka, a Human in the Loop).
Curious to test it out? Here's some examples of AI detectors you can try out for yourself: Copyleaks, Crossplag, Writer, and there are many more.
2) Lean into AI content for short form text
AI detectors need an average of 50+ words to detect AI generated content with accuracy. So, short form text like headlines, subject lines and even captions are more suitable for unaltered AI writing than long form blog posts.
3) Use AI as a brainstorming buddy
AI writing as a first draft > AI as a final product
Use AI as a brainstorming buddy to create outlines, organize thoughts, and even create a first draft that you plan to edit significantly.
4) Listen to your customers
What really matters with content is that it's useful to your end user. If you are using AI to create bad content, it's still going to be bad and unlikely to rank in Google or produce value for your bottom line.
Focus on creating valuable content, and value will follow.
p.s. We highly, highly recommend disclosing your use of AI in any content you create. Trust wins in business.
USE CASE DEEP DIVE
Google Sheets + GPT-3 use case: cleaning your messiest data
We've been using this GPT-3 Google Sheet extension to clean our transcription data.
Simply install the GPT for Sheets extension into a new Google Sheets file. (we shared a video step by step on TikTok)
Add your messy data
Write how you'd like your data cleaned in natural language
i.e. Our prompt was "Clean the following transcript by removing any Speaker N sections, time stamps, or items in brackets or <> to return one cohesive transcript in a single line"
Call GPT to clean your data using the =GPT() formula
And boom - watch your text data get cleaned in seconds!
You can even copy the generated content and "paste values only" into a new column to get the cleaned data without the formulas - you're welcome π
Note - we are not affiliated with GPT for Sheets. If you have questions about the extension's data privacy, please reach out to the creator. We recommend being cautious about putting any personal or private data in any AI tool.
LINKS
For your reading list π
What's going on in AI news...
Getty Images is suing the creators of Stable Diffusion for scraping its content -- if you missed it, we did a deep dive on the different perspectives of copyright
Microsoft announced general availability of Azure OpenAI Service as part of its commitment to democratizing AI
Our favorite listen of the week ...
Reid Hoffman spoke about the AI landscape, big tech and where opportunities lie. This talk is littered with great insights. Worth watching the whole recording.
And if you're really nerdy ...
Paper about HyDE: Hypothetical Document Embeddings - a trick for better LLM context retrieval
Paper about FiberAI, a deep learning model for automated analysis of DNA fibers
That's all!
We'll see you again on Tuesday. Thoughts, feedback and questions are much appreciated - respond here or shoot us a note at [email protected].
... and if someone forwarded this email to you, thank them π, and subscribe here!
Cheers,
πͺ The AI Exchange Team