Guide to Practical AI
The right tool now includes a growing suite of AI (Artificial Intelligence) image and writing tools. Groups like OpenAI.com have created tools that are now good enough to replicate the creative outputs of people. Scary: yes. Time-saving: also, yes. They are now becoming too good to ignore if you are in the creator space. In 2022, it was reported by just one AI tool built on OpenAI, that they created over 15bn words, for context, all of Wikipedia is 4bn.
Whether it’s graphic assets, video content, or a podcast, this guide will share the best way to DIY the multimedia needed.
The rise of chat-first search and the normalization of chat as a primary user interface has a lot of implications for nonprofits. In the near term, Whole Whale estimates that by 2025, chat-first search will eclipse traditional search leading to a steep drop in organic traffic and possibly threaten the Google Ad Grant. There will also be a growing need for organizations to develop their own chat tools trained on their data. For fun we have included our BETA chat tool just on this page in the bottom corner.
AI Text Generation & Search
Recent advances in natural language processing AI projects have finally gotten good enough to create original writing and imagery that mimics human output with amazing accuracy. Learning to work with these tools to increase the quality and quantity of creative work is a must. With the ability to browse and summarize the web, it is an incredible research partner.
Tools to check out:
- Chat.openai.com – you’d have to have been under a rock not to have played with this one
- Anthropic.com – an impressive rival to OpenAI, their AI “Claude” is impressive
- Hugging Face Chat – open source chat UI connecting to multiple models like LLAMA
- Cohere.ai – interesting to play with and has potential to be trained on your data more
- Bing Chat – basically GPT4 plus web browsing ability
- Google Bard – Google’s LLM in beta only
- Perplexity.ai – a GPT based search tool
AI Image Creation
A picture is worth a thousand words” is an age-old adage, but it’s probably right, give or take a few sentences. Now, you can actually turn those words into images thanks to generative AI projects like DALLE2 and Midjourney. There are tons of image editing tools out there, but the best one for novice designers is Canva—they provide great tutorials for quick self-teaching, and their nonprofit program that makes advanced features free. They have also integrated AI image generation. It is also an exciting time because of the amount of creativity that is now possible with image generators like DALL-E2 and MidJourney. Find our free course on using AI for creating copy and images.
Decide which tool is right for you:
- DALLE2 – Openai.com
- StableDiffusion – Open-source image generation
- MidJourney – Discord-based image generator
- Hugging Face – Open source AI image gen
- Imagen.research.google – Google’s Image generator
- Replicate.com – open models with various use-cases
- Autodraw.com – autocompletes your drawings, great for vectors
- Tome.app – creates presentations & storytelling decks
- Have a custom image creation page created: CauseWriter.AI
Once your team is through the experimentation phase, it will be important to create a branded image prompt that gives some visual consistency to images created across your site and socials. The new art of image prompt design will be one that key team members will need to figure out, here is one of the best image prompt libraries for inspiration.
AI voice replication, translation, and repair is being layered into every tool that deals with audio from podcasts to meeting recorders. For podcasts, it has never been easier to edit and create a high-quality audio project, here is our how-to guides for creating and measuring the success of your own podcast. For fun, our CEO created an Open AI GPT3 version of himself and then interviewed it this is no longer sci-fi…
Some of our favorite tools for AI Audio
- Eleven Labs – allows you to train a voice or create one.
- Descript.com – super easy text-based editing and AI audio repair
- Soundraw – create music on demand
- topmediai.com – quick way to turn text into celebrity audio
- murf.ai – great sounding voiceovers
AI video is under rapid development and the possibility for turning something like a pencil drawing into an animation will remove many of the technical roadblocks to creating quality video for organizations. There are also many tools able to create speaking avatars for transcribed text. These will all be evolving as more video data is trained into various LLMs.
Some tools to keep an eye on for AI Video:
- D-ID – avatar creation and narration
- RunwayML – text-to-video creator of short clips
- Kaiber.ai – turn a picture into a video
- Opus.pro – convert a long video into good clips
- Pictory.ai – can turn a script into great explainer videos with text overlays
- Descript.com – super easy text-based editing and connection to huge video library
- Replicate.com – experimental models
- Synthesia.io – creates talking avatars and videos based on text and presentation.
- Makeavideo.studio – creates short videos based on prompts
Created from “A whale jumping in a clear blue ocean”
The truth is there are more tools being created and improved per day than can be kept up with. Here is the best place to constantly shop for tools: Future Tools. This can be a wonderful way to waste time and go down random rabbit holes. Consider that 90% of these tools will be eliminated by feature updates from big players so have a bias toward the companies that you think will be around in 2 years. There is nothing sadder than training your team on a tool only to have it run out of funding.
Even current leaders like MidJourney are going to have to survive legal scrutiny from class action lawsuits around their use of artists’ images.
Whole Whale has developed CauseWriter.ai to custom-build internal chat tools to increase productivity trained on an organization’s own data.
Some tools as of Q2, 2023
- NVIDIA Canvas: https://www.nvidia.com/en-us/studio/c…
- Versy: https://www.versy.ai/
- Leia Pix: https://convert.leiapix.com/
- Pimeyes: https://pimeyes.com/en
- Deep Nostalgia: https://www.myheritage.com/deep-nosta…
- Luma Labs: https://lumalabs.ai/
- Reimagine Home: https://www.reimaginehome.ai/
- Do Not Pay: https://donotpay.com/
- Mubert: https://mubert.com/render/pricing?via…
- Eleven Labs: https://beta.elevenlabs.io/
- Mixo: https://mixo.io/?via=obscurious
- Looka: https://looka.com/
- Poly: https://poly.ai/
- Tome: https://beta.tome.app/
AI Ethics & Policy
Wait, if this is so important why’d you put it last… Because we knew the cool shiny parts above would bring readers in that would share this resource. Then, people like you would find this part and realize the real work that additionally needs to be done.
Just as a policy of rejection isn’t a viable strategy, neither is one of blind adoption. Consider where you sit, the power of position, and the advantages inherent of your organization (and self) before taking too hard of a stand either way.
As with all complex issues, leaders and teams must sort through their beliefs and evolve their policies. There are risks with data privacy, misinformation, and many other minefields yet to be revealed. ChatGPT was the fastest adoption of a product in human history, the chances that your team is already randomly using it is very high. The question is what is your policy and strategy for the conscious adoption of these powerful tools?