Aaron Swartz.
Activist, RSS 1.0 co-author and reddit co-founder, intensely believed in open free access to information. He downloaded millions of academic articles and research for the purpose of sharing information. The US promptly brought 13 federal felony charges, and he faced up to 35 years in federal prison. Legal pressure and depression made him commit suicide.
AI models are trained through ransacking infinitely more data, but nothing is done. Why rules for thee, but not for me?
What are datacenters? They're facilities that house the infrastructure and servers for, in our case, the running of AI models.
The issue though is how they work and what they cause. First, these servers need cooling, using up massive amounts of freshwater supply and polluting constantly. And second, their construction is destroying neighbourhoods, and sucking up the power grid, driving up electricity prices without creating permanent jobs. All without the approval of residents.
Some months ago, there was a surge on social media of deepfake NSFW content of women and their profiles via AI such as Grok on X.
This isn't anything completely new. However, that doesn't erase the fact that these virtual depictions of real people causes real harm and has very real consequences on their lives and mental health.
Do we really want an online space where anyone can have their image sullied like that by anybody?
Though steps have been taken to address this, it's not enough.
What are deepfakes?
Essentially, they're images or videos manipulated through Gen AI in which someone's likeness is being mimicked.
There's the immediate concern of misinformation arising, as this poses obvious qualms over public figures being made to say or do things that didn't happen or they don't condone.
However, more issues exist, such as the potential for scams and fraud to be committed with this tech, or the risk of having one's appearance denigrated and used to generate NSFW content.
Generative AI poses a hefty risk on the misinformation front as well.
Now that everyone can generate realistic videos and pictures of anyone who's ever shown themselves online willingly or not, we find ourselves in a scenario where every image or video online can be fake.
It's becoming increasingly common for duplicitous content involving public figures like politicians to show up from nowhere, feeding narratives and causing confusion.
Thus, it's crucial we learn to detect them as such
#NoAI
Some say Generative AI has "democratized" art. And while we get where these claims come from, we disagree vehemently.
First and foremost, the claim argues it makes art accessible to everybody. However, one has always been able to create with nothing but their hands, a pencil, or a sheet of paper. Humans have made art with nothing since painting in caves.
Also, the ableist angle used for the argument is, in itself, ableist through assuming disabled people can't make art on their own.
#NoAI
Remember those Ghibli-style AI generated images?
It might just be one of the most glaring examples of why "AI art" is an oxymoron.
Gen AI is trained to produce content with images and videos from all over the web, without asking for consent or crediting what it's been trained on, breaching copyright, for the sake of quick images.
Ghibli prides itself on their absurd levels of dedication, time, intent, and artistry put into the craft.
Mimicking their style is an insult to the industry.
#NoAI
Why select Generative AI as the main focus of our campaign?
As we said in the previous post, different types of AI exist, with different risks. With that in mind, we consider Generative AI to be the most harmful form of AI for society.
Our reasons stem from multiple concerns, such as deepfakes, misinformation, generation of NSFW content, copyright infringement, plagiarism, cognitive decline, and endangering artists.
Pandora's box opened, but we can take steps to mitigate the impact.
Is ALL AI bad?
Not necessarily! AI has its pros and cons, and what usage we give to these machines will dictate whether it's ultimately good or bad.
Likewise, not all types of AI are the same.
Analytical AI focuses on reviewing data to recognize patterns, make predictions, classify the data and reorganize it.
Generative AI, on the other hand, exists to produce content through training on existing content.
So while both use deep learning, they differ in purpose, effects, and ramifications.
Hi!
This is our communication campaign, in which we'll attempt to highlight numerous risks and dangers involving Generative AI, now more apparent than ever through its progression and usage in the last year.
Such concerns include what's most commonplace in talks about the hazardous nature of this tech, like qualms over "AI art", copyright issues, and deepfakes. However, that's not what we'll limit ourselves to during the duration of the campaign. 📢
Stay strong, keep holding the line! ❤️
#NoAI
