Responding to the two options for allowing AI training on my writing:
1. "the ability to opt out...but...a 10% boost in earnings on Medium to people who opted-in?"
My answer would depend partly on:
(a) Whether the training would be enabled per-story or per-author. I've published 430 stories on Medium. I might be willing to allow AI to train on some stories, but not others, for any number of reasons, especially based on what the stories are about and whether I feel they're personally or professionally significant.
(b) The amount of money on which I'd receive a 10% bonus. If I earn—hypothetically—$100/month across all 430 stories, allowing AI to train on my entire body of work for an extra $10/month is a "no."
(c) How long the AI training is ongoing. Once I give AI access to my work, wouldn't they just scrape my stories and then they'd have what they need? So they wouldn't pay me $10/month for the rest of my life, right?
(d) If Medium were to dissolve as a company, and the AI had already scraped my 430+ articles onto its system and continued to train on it, would the AI company legally owe me anything? There would be no Medium base earnings on which to calculate the 10%, and Medium would not exist to pay me.
2. "Would you allow a search engine to train on your writing in order to generate AI summarized answers that credited you?"
I'd be interested in Google directing more traffic to me and crediting me as an expert on a topic, of course. I'm imagining that would work like a Google "snippet" highlighted at the top of search results. However, as someone who writes on philosophically complex topics and is very picky about my exact word choice, I don't trust AI to summarize my writing, not this year and not next year—so an option that hangs on AI-written summaries is useless to me. However, in general, credit given to my name and/or more web traffic to me could be an acceptable form of compensation. If Google said simply at the top of its search results, "Tucker Lieberman appears to be an expert on this topic," that might have value for me. But having AI go on to paraphrase my views is not acceptable. An AI summary would serve Google by keeping their user a few seconds longer as they read the fake sentence, and it would provide less incentive for that person to go to Medium to check out what I really said.
