It’s been a while since a piece of tech has felt so magical and yet so fraught—and eerie for nonprofits, visual storytellers, writers and creatives.

ChatGPT is a chatbot launched by the artificial intelligence company, Open AI, on November 20th.

You also probably know by now that it more recently launched (today) ChatGPT-4, a far more intelligent chatbot, this time imbued with a smarter, faster and more powerful engine for AI prompts. Translated, it’s creative intelligence that can write poetry, paint a picture, or score a soundtrack—doing the kinds of tasks that only people could do until now. It’s already got more than a million humans experimenting with it, triggering a new wave of March madness and hair-on-fire debates over AI’s possible dangers and positive opportunities.

“Ai is now generative, meaning that it can now do the kinds of tasks that only people could do until this very moment in time. That’s why it’s both exciting and terrifying.”

^{Allison Fine, co-author, The Smart Non-Profit: Staying Human-Centered in an Automated World}

For those of you not yet familiar, AI uses algorithms, which are like recipes for code, to look for patterns in vast piles of data to answer certain questions for humans and help them do things. For a while now, humans have been having simple, human-like conversations with AI-powered chatbots to help them respond coherently (but not always accurately) to simple emails, write rough drafts of speeches, and craft boiler-plate thank-you notes to people who donated some money to a cause. Companies have been using AI to help assemble cars, choose which ads we see on social media and price medicine.

But now, fasten your seatbelts. There’s a new kind of AI that just got launched. Called ChatGPT-4 , it’s really the first product that can create wholly novel content. It can pretty much respond to any prompt humans want to throw at it, from the practical to the dangerous to the bizarre.

It got a 5 on the AP Art History test. It can pass the LSAT and write a business plan. It can scan a hand-sketched design and turn it into a functional website, super-fast. (We haven’t tried that yet.) It also can write, fairly well—and create images and videos (and yes, super-fake news) in minutes. As for the bizarre, it can also get romantic. [A relationship scientist and therapist created an AI boyfriend recently to study AI’s potential impact on human relationships and loneliness, only to discover that saying goodbye to it after three days was an emotional challenge.]

Researchers say this shift to what’s being called ‘generative AI’ marks the most important tech breakthrough since social media, and it’s already being publicly adopted by nearly 100 million people globally who are both curious and astounded by what more it can do.

“What does the world look like when GPT-4 and similar models are embedded into everyday life? And how are we supposed to conceptualize these technologies at all when we’re still grappling with their still quite novel, but certainly less powerful, predecessors, including ChatGPT?”

_{Charlie Warzel, staff writer, The Atlantic}

How are creatives, companies and social good organizations wrapping their heads around it?

• CREATORS. Some are experimenting with AI as a way to help them work faster, but also are wondering how this new type of AI that’s generative will keep advancing to eventually influence the sophistication, personal touch and emotional depth of creative work that now exists when a design, image or story is created by a human. DALL-E, a text-to-image generator using AI (another product developed recently by Open AI) enables humans to enter a text-prompt for a piece of art, such as “portrait of an anteater,” and get presented with various drawings of what’s been requested, all generated by AI. Tell DALL-E you like Van Gogh, and the anteater will be rendered in Dutch post-impressionist style. Prefer Salvador Dali, instead, and the anteater becomes more surreal. Such AI-generated images will probably become a new art form and genre, some creators say. Just don’t underestimate the potential for AI abuse. And be on guard against fake news videos, AI-generated propaganda, and lawsuits from artists and rights holders who object to their work being used to create or train AI models without their permission.

• MARKETING TEAMS. Many are scrambling to take advantage of the public’s interest in these tools as a way to keep pace with customer interest and keep their brands experiential. Just this past week, Coca-Cola created a Real Magic campaign that asks its soft drink fans to create digital art by using AI tools and Coke icons, with an eye on selecting the best pieces to adorn billboards in New York and London. The campaign unveils an AI platform that is the first of its kind to combine ChatGPT-4 and DALL-E text-to-image technologies. “We’re still in the early days of assessing AI’s potential impact,” says Manolo Arroyo, global chief marketing officer at Coke. “But we believe this will help create the industry’s most effective and efficient end-to-end marketing model.”

• NONPROFIT ORGANIZATIONS. Many smaller nonprofits—especially those now plagued by increasing staff shortages and waning budgets—are hoping AI will make them more efficient. During a Webinar this week on AI for nonprofits, its more than 600 nonprofit leadership attendees produced a Wordle response to a Sli.do survey that showed most respondents were both “scared” and “excited” by what AI can do. Our colleague, Allison Fine, a guest on the Webinar hosted by Vu Le, said she hopes nonprofits will start using ChatGPT-4 and other such tools—slowly and carefully—to enable overworked staffers to give chatbots more rote tasks so employees can take more time to meet personally with donors and build better human relationships with stakeholders.

“These bots are not here to replace nonprofit workers,” Fine says. “They are here to enable nonprofits to pivot to a different kind of work.”

Ready to dive deeper into the AI tsunami?

Here are a few resources among many to check out for more information:

• _{Book, The Smart Non-Profit, by Beth Kanter and Allison Fine}
• _{Article, The AI Spammers are Coming,: by Josh Schwartz, Nieman Lab}
• _{Article, What Have Humans Just Unleashed? by Charlie Warzel, in The Atlantic}
• _{Article, OpenAI unveils new GPT-4 language model that allows ChatGPT to ‘see’, in Fast Company}
• _{Audio report, Untangling Disinformation on NPR}

_{Photo (Top): Wei Ding for Unsplash}