Dear Matt. 404 Media tells me you're in the process of selling access to the data I've published on WordPress.com and Tumblr to the AI companies OpenAI and Midjourney and that I have to actively opt-out if I don't want my data included.
I have ten questions for you regarding this that I think most of your users would also like to hear your answers to:
- Why do users have to opt-out of sharing their data with AI companies? This assumes most users agree to share their content with AI services. What data and/or reasoning backs up this assumption?
- Who gets the revenues from this data sharing, and how much are those revenues? Specifically, are creators being compensated for their data being sold to a third party?
- Who decides on behalf of abandoned sites or sites whose creators are no longer with us? Not everyone has the capability of opting out. Who speaks on their behalf and protects their interests?
- How were the affected users consulted on this? And what was their feedback?
- What professionals were consulted, and what did they say? Did you consult with legal? Did you consult with your ethics officer (I'm assuming you have one)? Who else were involved in this decision?
- How does selling this data line up with your open source principle of users owning their own content? Who do you believe own and have the right to profit from user content hosted on your platforms?
- Is the data being sold only from free sites, or does it also include sites the user pays you to host on your platform? If the latter, how do you justify "double-dipping" in the revenue stream?
- Why do you believe this is the right decision to make on behalf of your users? And how do you respond to those who say it is not?
- How did you pick these commercial AI companies over open source alternatives? Reporting indicates you are in talks with OpenAI and Midjourney. Neither is open source.
- Why should we trust you with our data going forward? See The Five Questions of Tony Benn.
Please respond with a post of your own and link it back here.
Originally posted on mor10.com
Creates content to fuel curiousity
9 个月I haven't been a part of the Wordpress community for a while now, but the hubris involved in this is incredible. I somewhat understand trying to find revenue streams for supporting free social media products like Tumblr, but selling data that creatives have already paid you to host is beyond bonkers and highly unethical. At least not without offering a way to compensate people for their contributions. Offering a financial incentive to users to do something you want them to do (like watch ads) is fine, but they probably won't offer a cheaper plan that would allow for AI training because they know their core users would hate that.
UK tech policy wonk
9 个月Here are some of your answers, Morten: https://www.404media.co/wordpress-firehose-allows-ai-companies-to-buy-access-to-a-million-posts-a-day/
Director of Standfirst, Interconnect and Design Week.
9 个月And let's not forget the WordPress Firehose that by default sells data from your self-hosted blog if you happen to install JetPack.
UK tech policy wonk
9 个月The 404 story and podcast cited the fact that these leaks were a direct consequence of a completely separate story, which was Matt's public behaviour last week towards a community member, including cross-platform stalking, harassment, doxxing, and kink-shaming, all facilitated by inappropriate access to internal system data. It must be said that none of that behaviour was news to anyone driven out of the WP community, including you and me; the rest of the world just sees it now. One member of the trust and safety profession (link below) cited his professional misconduct as the worst thing she'd seen in her career on the internet: which is quite something coming from someone who spends all day looking at the worst things on the internet. As she put it, "it’s likely no coincidence these internal documents were leaked so close to [his] behavior last week - it indicates to me ongoing internal strife exacerbated by [his] misconduct." Given all that and all it implies, there is just cause for an 11th question to your list, related to the 10th but more to the point. I'll leave it to your readers to suss it out. https://www.supporthuman.cx/25feb24-roundup-a-masterclass-on-why-we-have-trust-safety-teams/