The more I hear about AI learning to write essays to make art to sell at fancy auction houses, the more I want to know why AI isn’t washing my dishes are doing my laundry or changing my oil. We don’t need AI to make art. If we need it at all (highly debatable) we need it to do the mundane things so we can all spent our time making the art.

AI rapidly becoming mainstream, and in many cases the perception is that it’s innocuous. But the reality is very different, and here are some concrete reasons we will not be using AI to create brands or copy or art or websites.

1. AI is being used to enact genocide in Palestine

We might as well start with the biggie, because the rest hardly matter in the face of this one. AI is being used by global powers to commit genocide agains the Palestinian people. Google and Amazon have provided AI tools to the Israeli military. These tools are used to analyze data and select military target. The Israeli Defence Force also developed its own AI tools to called Lavender, used to select targets and which users have reported had built-in allowances for civilian casualties. A lot of things originally developed for the military have become useful mainstream products—from duct tape and micro waves to the internet itself—but AI’s origins as a tool for mass death should be reason enough to keep it out of the mainstream (or at least our own personal Flegg Creative mainstream).

2. AI is contributing to climate change

There are all sorts of wild claims about this one, like ChatGPT incinerates a bunny every time someone asks it to write an email, but here’s the thing: even if we disregard the brashest claims, the most dramatic, the most difficult to quantify or verify, it’s still bad. While it might sound counterintuitive to talk about the internet—this magical, intangible, invisible-but-everywhere global cloud—burning fossils fuels, but the reality is that the magical cloud relies on real, physical servers and data centres to store and transmit the information available at all of our fingertips. The more we use the internet, and more specifically the more we create data, the greater the need for huge data centres housing tens of thousands of servers all requiring power 24/7/365. In addition to the power to keep them running, these servers also require water to cool them down. A huge amount of our electricity comes from fossil fuels, hence the use of AI resulting directly in increased carbon emissions.

Training AI models, especially complex language model, is extremely energy intensive, and while that level of consumption doesn’t last forever (at least theoretically) the continuous adaptation of these tools means they will continue learning and changing and therefore using huge amounts of power and water well into the future. So when OpenAI CEO Sam Altman says that thanking ChatGPT costs him millions of dollars, he is literally referring to the cost of energy and water.

3. AI is trained on stolen art

Remember when we were teenagers and threatened with imprisonment for downloading Avril Lavigne songs on Limewire? Or when we were in university and had the fear of god put in us about plagiarism, convinced that a misplaced colon in a footnote would lead to a failing grade, expulsion and (probably) imprisonment? Well, no more. Now AI gets free access to all kinds of artwork thanks to datasets created to teach AI to write its own academic essays, craft novels and screenplays and create digital art.

A CBC News investigation discovered that a dataset containing 2500 copyrighted works by 1200 Canadian writers had been used to train various AI platforms, including OpenAI and ChatGPT. This Los Angeles Times piece links to a host of other lawsuits launched by writers, visual artists, actors and audio creators against a variety of generative AI companies for the theft of their work. Artists (if they’re lucky) earn money through royalties, licensing fees and commissions, and the use of this copyrighted material without consent or payment is just theft. The fact that it’s multi-billion-dollar companies doing the stealing just adds insult to injury.

4. AI is unreliable

Because AI, especially the bit that now appear at the top of every Google search, pulls from all over the internet, and the internet is full of inaccuracies and omissions and downright fabrications, those summaries are often full of inaccuracies, omissions and fabrications. Janel Comeau, writer for noted Canadian comedy news site The Beaverton, did some internet searching and discovered that both Meta and Google AI engines think Cape Breton is in its own time zone, 12 minutes ahead of the rest of Nova Scotia, citing a Beaverton article she wrote.

A friend recently shared this completely bonkers story that illustrates how, apparently, we’ve done a better job teaching AI to lie to our faces than we have teaching it basic reading comprehension. Writer Amanda Guinzburg asked ChatGPT to help here choose a collection of her essays to include with a query letter. The bot goes on to completely make up the summaries citing topics and even wording that appears nowhere in any of Guinzburg’s work, admit that it lied, lie again and then finally admit that the format of some of the essays is not actually legible to it at all. Give it a read. It’s truly wild.

5. We believe in people

When it comes to our own work, the human part of human-centred design is pretty key to the whole thing. Call us old-fashioned, or corny or cringe (one of us is cool and the other is pretty cringe, I’ll leave it to you to figure out who is who) but we believe that real people create beautiful, unique, useful, unexpected things, and that is the whole joy of it. As challenging as working with humans can be, the magic happens at the intersection of expertise and experience that comes when clients share their vision with us and we apply our skills in service of bringing that vision to life.

We want to make meaningful art and writing, and create designs that convey real messages and connect with real people. And we want to support you to do the same. AI only undermines those goals, and so this is our promise that working with us means working with people, plain and simple, in all our imperfect glory.