Bhagpuss published a post a few weeks back about AI and how we feel about it. Not about the hard crunchy aspects of AI, but about feelings and attitudes towards what may one day be a new form of actual intelligence we have to share the world with. I have some thoughts on the topic that I could either make into a long comment or a short post, so here we are.
Feelings
My take on Bhagpuss’ central thesis is that, if we divorce ourselves from all of the systemic harms that AI can and does cause, we are left with what might be a new form of intelligence that we have to decide how to deal with. And his position is that we should approach that intelligence with kindness instead of cruelty. We shouldn’t be lining up to smash the AI “Box” when it comes to our homes but should instead embrace it as a new kind of friend.
I have two categories of thought regarding AI harms.
- The harms caused by the creation and operation of AI: think data centres, power generation smoke stacks, and infrastructure consumption. I wrote about these things to a goodly extent in a previous post and don’t need to repeat that here other than to state that these expenditures apply to all large language model AI whether the resulting model runs stand-alone in a mobile device or in “the cloud”. The model must still be generated, and that is what uses the majority of data centre computing resources. I didn’t touch on the consumption of human-generated content and the subversion of content ownership that AI thrives on, but that fits in this category as well
- The harms caused through the use of AI: this includes loss of jobs, AI-caused deaths and injury in automobile accidents, and the use of AI in weapons platforms. It also covers AI supported mass surveillance and things like AI driven botnets attacking critical services. How humans use AI won’t be warm and fuzzy because it is in the cold and spikey edges of such use that profit is maximized.
So, considering the harms that AI can and does cause, do I “hate” artificial intelligence? No, I do not. If I can get an AI agent that makes my life easier or happier without simultaneously boiling an ocean somewhere, chances are I’ll use it. I will consider the harms and weigh them before I make my choice. And I firmly believe that there need to be regulations on how AI is developed and how it is used.
Where I go from here
I have to keep paying attention to how AI is developing and what it is being used for. I will always advocate for good, common-sense regulation of large AI companies and of the government itself. I won’t want to trust the AI companies to self regulate, and I will question any laws they lobby for themselves. I’m also going to be skeptical of how governments use AI in surveillance and military applications.
I won’t hate AI itself, at least not until the rise of Skynet, but rather how it is created and used. I will continue to use AI for the things it is helpful to me for, while being cognizant of the occasional hypocrisy such use will undoubtedly lead me to. I just hope I’m not going to be living in a post-apocalyptic hellscape wrought by mistakes in judgement about the risks that AI poses.
The header image for this post is AI generated by ChatGPT. I was running short of free images so it doesn’t do a very good job depicting the topic of this post, but it is close enough for AI