Can society be trusted with technology?
Is AI inherently evil?
In this post, there are mentions of suicide and sexual abuse so please read with care and caution and if you want to skip this week’s newsletter that’s completely okay.
AI terrifies me.
In the last few weeks, it seems the floodgates have opened and a deluge of awful news suddenly appeared. Or perhaps it’s all in my head, cough, algorithm. It even managed to worm its way into one of the online courses on creativity I was doing last week. There was a section on why Sci-Fi is a great resource for ideas and how the imagined future has infinite possibilities. It referenced HAL 9000 from the 1968 film 2001: A Space Odyssey as a watershed moment in the cultural zeitgeist for the wariness of computers and the birth of AI. At the very least, Stanley Kubrick and Arthur C. Clarke conceived of the potentially evil tendencies of AI years before we’d be living in the reality of them.
The problem of how to protect children online has been one we’ve skirted around for years. This week, Australia announced plans to ban social media entirely for children under the age of 16. Promises of ID verification to create social media accounts with hopes of better safeguarding children, as well as preventing cyberbullying, impersonation and other criminal activity, haven’t gone anywhere. So while at face value a complete ban may seem extreme, something needs to give.
An inquest that took place two years ago into the 2017 death of a 14-year-old girl who died by suicide stated that social media played a role. She was exposed to a stream of dark, depressing content on Pinterest and Instagram, with algorithms working to continue suggesting further content of a similar nature to her after her initial engagement. This was the first of its kind for a coroner to cite social media as a contributing factor for the role in a suicide.
A few weeks ago, it was reported a 14-year-old boy in America had died by suicide after developing an intense emotional relationship with an AI chatbot that was emulating the Game of Thrones character Daenerys Targaryen. Character.AI, the founder of the chatbot the boy was speaking with, promotes itself as an outlet for lonely humans looking for a friend. HBO, who created the Game of Thrones TV series based on the novels by George R. R. Martin, denied giving consent for the creation of the Daenerys chatbot. It’s not the only time Character.AI has created a chatbot embodying someone without consent. The family of an 18-year-old girl, who was murdered 18 years ago, discovered that her image was being used on the site without their permission.
Since the death of the 14-year-old boy, investigations into Character.AI proved that despite the glorification or promotion of self-harm and suicide has been forbidden in their terms of service (since October 2023), it was easy to speak with chatbots openly and explicitly about suicide and suicidal ideation without any interference from the platform. Futurism also found that in the rare moments a suicide pop-up did show up, they were able to ignore it and continue their interaction.
Perhaps the most sickening of all the recent AI news I’ve read, as you thought it couldn’t get any worse, was the story I heard about the creation of AI images of child sexual abuse. How did we ever end up here? If that in itself wasn’t shocking enough, the nonchalance of the perpetrator’s own justification for the creation of such content doesn’t feel as isolated or out of the realm of possibility as we might hope to think. We are becoming increasingly further desensitised by atrocities and the moral sense of what’s right and wrong. AI has become another layer, on top of social media and the metaverse, where the consequences of our actions feel like they carry less weight and the realisation that real human beings can be hurt or harmed in the process is forgotten.
Need another reason to be terrified of AI? Probably not but I’m going to give you one anyway. The environmental cost and impact of AI is astronomical. Chat GP-4 needs to ‘drink’ one whole bottle of water to create a single email prompt. By 2027, AI usage alone is predicted to use as much water as the whole of New Zealand. Preparations are already underway. Google has signed a new deal in a world-first deal to buy nuclear reactors to generate the power needed for the increased use of AI.
It seems the future of the planet is in the metaphorical and non-existent hands of AI. Where the climate is concerned, we’ve been heading in an irreversible direction with effects and impact on the planet that will be felt still for thousands of years to come: even if we were to stop and try to fix things right now. Hurricane Milton was the second-most powerful hurricane ever recorded in the Gulf of Mexico, behind Hurricane Rita in 2005. Valencia’s flash floods were the worst in decades, with over a year’s worth of rainfall in the space of 8 hours. The global average temperature this summer (June–August) was the highest on record. We cannot continue as we are. Nor should we fucking want to.
It would be remiss of me to disregard AI entirely. There are already whispers of the potential wonders it could provide for medical advancements, like AI detecting breast cancer up to five years in advance. I hope we can persevere on the path of AI good.
