Using AI for customer service is the stuff of my nightmares. link to ‘Zoom will soon integrate Anthropic’s chatbot across its platform - The Verge’

Some good comments in here—especially on how AI enforces and normalizes certain kinds of writing instead of allowing us to determine what writing should look like. link to ‘Google’s AI pitch is a recipe for email hell - The Verge’

Content moderation is hard, and it’s especially hard at scale. Because AI makes doing things at scale easier, it necessarily makes content moderation harder. link to ‘Spotify ejects thousands of AI-made songs in purge of fake streams | Ars Technica’

Very interesting look at some of the less obvious implications of generative AI. link to ‘People are arguing in court that real images are deepfakes : NPR’

Heading into finals, campus sent out a message about AI detection tools maybe not being trustworthy, which is great. However, this is in the context of these tools being wrapped into plagiarism detection software we already have access to, so they should say the same about it, too.

This is one of the most amazing things I’ve read on generative AI. link to ‘Bing has a testimony of the Book of Mormon! And other adventures with AI chatbots.’

I am not an AI expert, and my concerns aren’t on the existential scale. However, I do think it’s important to avoid moving fast and breaking things with these powerful technologies. That isn’t necessarily to say that more powerful AI shouldn’t be released (though I’m already disinterested by the current stuff), just that racing to improve them for commercial benefit and as technological flourish doesn’t strike me as socially responsible. link to ‘In Sudden Alarm, Tech Doyens Call for a Pause on ChatGPT | WIRED’

Lots of helpful stuff in here. link to ‘ChatGPT Is So Bad at Essays That Professors Can Spot It Instantly’

Very interesting! I know some critics will describe this as a morally panicked response, but I disagree. I think it’s smart to ask how AI will affect human creators and for companies/communities like Paizo to take principled stances. link to ‘Paizo bans AI-generated content to support ‘human professionals’ - The Verge’

This is dumb and worrying. The CEO of Gab has been promising to develop “based AI,” but he’s a bit player. Musk has the resources and influence to make this a bigger problem. link to ‘Elon Musk Is Reportedly Building ‘Based AI’ Because ChatGPT Is Too Woke’

This is my third time reading this book—I couldn’t resist coming back to it for the “epistolary novel” square of my library’s “Books and Bites Bingo” challenge this year. The print book is amazing, the audiobook manages to adapt a book that shouldn’t be adaptable, and I enjoyed this read as much as the last two. The language and worldbuilding are subtle but effective, it’s morally complex without trying too hard to be, and the characters are a good mix between believable and, well, archetypal characters in a YA novel.

Interesting and important read. link to ‘Voice Actors Push Back Against Their Voices Being Used by AI’

Why… why don’t we better anticipate better misuses like this? Are technological “progress” and market opportunities more important than these side effects? link to ‘4chan users embrace AI voice clone tool to generate celebrity hatespeech - The Verge’

Here, as with autocorrect and citation managers, my personal opinion is that any human who knows enough to use the tool critically knows enough to do the job themself. Maybe slower, sure, but slower isn’t always bad. link to ‘CNET Defends Use of AI Blogger After Embarrassing 163-Word Correction: ‘Humans Make Mistakes, Too’’

Fascinating read on potential threats posed by AI—and potential solutions. link to ‘How ‘radioactive data’ could help reveal malicious AIs - The Verge’

Well, this is terrifying. link to ‘Experts Warn ChatGPT Could Democratize Cybercrime - Infosecurity Magazine’

Good thing engineers really anticipated and considered these consequences before developing this software, right? link to ‘Thanks to AI, it’s probably time to take your photos off the Internet | Ars Technica’

A helpful and thoughtful critique of how people are doing AI text generation. link to ‘ChatGPT, Galactica, and the Progress Trap | WIRED’

Very interesting read. link to ‘Facebook Pulls Its New ‘AI For Science’ Because It’s Broken and Terrible’

Really important story here, and glad to see George Veletsianos quoted. I’ve long been an advocate for developing assessments that are impossible to cheat at, but I don’t know if that’s the entire (or even a practical) response to GPT-3. We are continuing to develop technologies whose societal effects we are not prepares for. link to ‘Students Are Using AI to Write Their Papers, Because Of Course They Are’

Ooof. AI-generated art is fun, but it comes at a price, and we can’t afford to forget it. link to ‘AI Is Probably Using Your Images and It’s Not Easy to Opt Out’

Compelling podcast episode from Mozilla highlighting morally dubious uses of AI. It’s really important that we be more reflective about this instead of trying things and seeing where they lead. link to ‘The Tech We Won’t Build — The Internet Health Report 2022’

It amazes me that we all agree autocorrect spectacularly and regularly fails and yet believe that we can trust fundamentally the same tech to do harder work like grade homework, flag content, and suggest prison sentences.

Yet another Grammarly ad has me thinking… I think my skepticism about AI is not so much the mistakes it makes so much as it is the assumption that human experiences are so well-structured that they can be reduced to an algorithm.