The Dangers of AI
Introduction
Recently, there’s been an increasing trend where people are replacing their social media profile pictures with those they have generated using Chat GPT or other AI tools. AI is becoming increasingly normalized in ways we are barely able to understand or recognize; it offers to summarize our emails or PDFs, the AI assistant often shows up in Zoom meetings to “take notes,” and a summary shows up with every Google search. While it may seem like a helpful new technological tool, there’s a lot of dangers of AI and ChatGPT that people need to be aware of. In this blog, I’ll delve into some of these dangers, particularly as they relate to AI being dehumanizing, significant privacy and safety concerns, the negative impacts for creatives, and the environmental impact.
How AI Shapes Our Human Experience
AI is ultimately dehumanizing. It is continually shaping and changing our human experience. Many people are now turning to Chat GPT for therapy, which is incredibly dangerous and removes the critical human element of therapy. So much of our experience as humans is shaped by community, connection, and creativity, which AI is continually attempting to infringe on. That makes it even more important for us to resist this. For instance, rather than turning to AI for answers on best ways to cook a certain dish, we can ask a friend or family member with experience cooking that. When recovering from surgery, we can gather advice from people who have gone through the same or similar surgeries in the past. And we can go to professionals (licensed therapists) for therapy and seek community connection over AI chatbots.
There’s also the issue that AI is often wrong, and is contributing to the rising issue of misinformation. While sometimes the incorrect things, often termed “AI hallucinations”, can be amusing, this is also dangerous. AI can generate fake images, invent celebrity hoaxes, spread false information about health and safety, and ultimately, spread lies. While we should always be wary of what we read online, and look for reputable sources, the proliferation of AI use has made that exceptionally harder, with many people just believing whatever they read.
The final issue I want to address is that AI is causing us to lose critical skills. When we believe everything we read online, we aren’t engaging with our media in a critical, evaluative way. When we outsource our thinking to AI, we lose our problem-solving skills. When we ask AI to summarize emails and documents, we aren’t spending the time we should to read and summarize these ourselves. When we ask AI to write or “clean up” our writing, we lose our writing skills. When we ask AI for ideas for a project, we are de-valuing our inherent creativity. AI is literally causing us to lose our skills, abilities, and creativity.
AI and Privacy, Safety, and Violence
When someone uses AI to generate a profile picture, AI takes their original image or images to make a new one. What happens with that picture after? It gets stored in their databases to be able to generate other images and photos. Similarly, when they are summarizing our emails and our Zoom meetings, where does the information from that go? From our web searches? Every prompt that is put in, every summary, is stored in their databases. Most AI systems have very vague or inconsistent privacy policies, which means that private information – our faces, our documents, our location – is stored in systems we don’t have any understanding of. This information is used to train AI, and they can alter such information however they choose.
The threat of cybersecurity is a genuine one. When AI is able to gather all of this data about us – our location, our work, our family – then it makes it much easier to use this as a weapon. In fact, only about 24% of AI projects have components to ensure cybersecurity. This leaves people and businesses open to scammers, hackers, and cyberattacks. AI has also been used to “dox” people, particularly those from equity-deserving communities, by revealing their schools, places of work, and even home addresses, placing even more vulnerability on communities that are often targeted by violence.
We are already seeing how AI is both creating new forms of abuse and amplifying existing types of abuse for communities most vulnerable, particularly women. AI is enabling image-based abuse through deepfakes, which is when images of people (particularly women and children) are turned into pornographic pictures and videos. This is nonconsensual and is deeply dangerous. This also increases the risk of blackmail and risk of catfishing, further endangering populations who are already vulnerable to sexual violence. It’s been shown that deepfake videos online in 2023 increased by 550% compared to 2019, and this number is expected to rise as the use of AI becomes more prevalent. Ultimately, choosing to use AI, even for something as simple as generating a new profile picture or taking notes during a Zoom meeting, exposes people and businesses to serious privacy breaches, safety concerns, and the potential for people seeking to do harm to use AI to abuse others.
Negative Impacts for Creatives
Just as AI takes our data to train its programs, it has been well documented that AI has stolen work from creatives: paintings, photographs, poems, academic articles, music, books. They claim stealing such work is necessary for “training” their programs, but this is ultimately done without the consent or compensation of the original creators of the work. Their programs are trained off of human creativity, where they proceed to steal that work to create “new”, non-human work. This has serious implications for those of us in the creative field.
One such negative impact is the loss of income and opportunity for creatives. Those of us who create art (I use this term broadly) are not compensated when AI steals our work to train its programs. People who generate images and content using our stolen work don’t compensate us for this stolen work, and may not even realize or care that it has been stolen. And when people can generate their own images, music, and books (albeit crappy versions), they end up not commissioning and hiring actual creatives – those of us who have spent time, money, and energy to learn our craft(s) and become good at what we do. It’s estimated that creatives are at risk of losing over 20% of income as a result of AI. This is different from someone learning how to design a poster themselves so they don’t have to hire someone to do it – AI blends the work of people who have previously designed posters to spit out a stolen design.
Another significant impact for creatives is that we need to make our stance on AI clear. People who are against AI, like myself, are more likely to question creatives as to whether or not AI was used in someone’s creative process, whether that was to generate an image, to brainstorm ideas for a book, or a backing track for a musical composition. It is important to ask these questions, to find out from creatives if there was any AI usage so that we can each make informed decisions about our own ethics. Personally, I will never financially or otherwise support someone who uses AI, for all the reasons I outline in this blog. However, this does place the onus on creatives to make our stance on AI explicit.
Unfortunately, there are far too many “creatives” who both hide that they use AI, and attempt to profit off of it. As a result, there has been increased mistrust of creatives, where creatives – even those of us who have made explicit statements against AI – will be accused of using AI. There are claims that there are certain “tells” of AI use, most of which are actually just signs of strong writing (you can pry em-dashes from my cold, dead hands). Again, AI was trained off of those of us who are professional writers and creatives, so unfortunately AI can sometimes reflect it. But for those of us whose creative work is 100% human generated, it can be frustrating feeling like we have to “prove” that it wasn’t AI-generated, while knowing that AI has stolen our work and the other work of creatives.
Environmental Costs of AI
A significantly concerning danger of AI is the environmental cost. This includes concerns for the amount of raw materials needed, the electronic waste, and the substantial amount of water and energy that is used. The increasing use of AI has caused the number of data centres to grow to 8 million, and this is expected to continue, causing even further damage to the environment.
Some of the ways in which AI is particularly detrimental to the environment include:
Microchips that power AI require rare earth elements, which are mined in ways that are extremely destructive to the environment.
The increased production of electronic waste results in hazardous substances like mercury and lead being put into the environment.
A request made through Chat GPT consumes 10 times the amount of electricity compared to a Google search. Such electricity usage results in burning of fossil fuels.
AI is expected to put 22 to 44 million metric tonnes of carbon dioxide into the atmosphere by 2030 (the equivalent of 5 to 10 million cars added to the road)
Data centers use water for construction and cooling electrical components. It’s estimated that AI centers consume six times more water than Denmark (population 6 million), and may drain between 700 and 1,100 million cubic meters of water per year (annual household usage of 6 to 10 million people).
Another issue relating to the environment is that AI is fueling the historical pattern of environmental racism. Environmental racism refers to how marginalized communities, particularly with high populations of racialized people and/or low-income people, bear the negative outcomes of environmental pollution. This is being translated into AI because many data centers are located in predominantly racialized communities, such as one in Boxtown, which is 90% Black with a median household income of $36,000. The community is experiencing health effects as a result of the increased level of pollution generated from the data center. When the environment is endangered, our health as communities – especially marginalized communities – is further at risk.
Ultimately, given how much the environment is already in danger because of human-made climate change, we absolutely cannot afford to support a system that is so immensely dangerous to the environment. The global temperature is expected to rise 1.5 to 2 degrees Celsius within the century, unless serious changes are made. With just one degree rise, we have seen increased natural disasters, climate refugees, extinctions and endangering of species and ecosystems, rising sea levels, and more – and that’s just here in Canada. We are at a pivotal moment where we need to be seeking solutions for protecting the environment and addressing the climate crisis, not further exacerbating it.
Conclusion
Knowing all of this, what do we do? Given how common AI has become, how do we resist? Here are a few strategies.
If you notice friends, family, or coworkers using AI – whether for profile pictures, posters, work documents, or any reason – invite them into a conversation about the harms of AI use. Feel free to share this blog and other resources. The more we can educate people and bring them on board with resisting AI, the better.
If creatives that you follow/engage with haven’t made an explicit statement against AI, ask them to do so.
Don’t engage with any AI content – don’t give it a “like”, don’t follow the creator, don’t share content. Ignore AI summaries on email and search engines. Be on the lookout for AI content, but don’t fall into the trap of assuming something that’s well written is AI (we’re just good writers).
If your workplace, school, or organization is encouraging the use of AI, help to educate them and ask them to create/adopt a policy against AI use.
Make sure your healthcare team (doctors, therapists, etc) are not using AI – this is major issue of privacy violations and they should not be doing this at all, but especially not without your consent.
Advocate with your elected representatives for policies about AI, particularly focused on privacy and safety. For instance, you can sign Petition e-7002 (Canada).
Every time that you engage with AI or generate something, you are implicitly endorsing all of these deeply harmful things that I’ve discussed in this blog (and there are many more that I haven’t gotten to). As consumers, creatives, and people, we have an opportunity and responsibility to make conscious, ethical choices – for ourselves, for our communities, and for our environments.
Take a stand against AI today. Stand with your community. Stand with your creatives. Stand with our environment. Stand for human connection, creativity, and community.
