The idea is to have a mix. In the same Brief page, there is already an Explainer section, which is supposed to be as neutral as possible, but there are also other sections (the Contentious Points section, Common Questions, Q&A, etc) where users' different opinions and takes can emerge.
For example, let's imagine a hypothetical Brief on the OpenAI/ Hugging Face incident. In the Q&A section of this Brief, creator and journalist users can post questions like "How bad is this, really"? and AI safety researchers and orgs can post answers. The answers, as you can imagine, can turn out to be more like takes, so not very neutral at all.
in short, the idea is to have the possibility of both learning about the facts in an unbiased way, but in parallel also have access to all of the different takes, disagreements, etc on a particular topic or event.
The video walkthrough (like in the article) shows more details on this.
Let me know if you have any more questions or suggestions :)
Thanks so much for writing about this. I also strongly believe that there should be a lot more effort going in AI safety communications. I'd like to share a few things: 1. My own recent project (just launched, also with an EA Forum post): telltheworld.io. I'm looking for all the feedback I can get, so it'd be great to get in touch or get warm intros with anyone in your network who could find this useful. 2. Another new AI safety comms project: https://commonsignals.org/ 3. The Effective Communications WhatsApp group. If you're not already on it, you might want to join - lots of AI safety comms people there, which I believe share your same concerns.
Hi Karen, I'm working on a new comms project in this exact direction, aimed at bringing more creators and journalists to AI safety communications. I just posted about it a few days ago, here on the EA Forum. Would love to have your take on it, if it's of any interested (not sure I can post a direct link in here, as I'm relatively new to the Forum).
Hey, thanks for the link. I know there are already projects automating this kind of summarisation work. However, that is not what I'm going for with Tell The World.
The goal with Tell The World is not only to make this kind of material more easily digestible, but also to have creators and journalists be able to confidently and credibly deliver it to their audiences. That is only possible if sources come with a "stamp of approval" by human experts - which is what this project provides.
Another major goal is to get more creators and journalists outside of the AI safety space to learn about & cover AI catstrophic risk, and - why not - become part of a larger community aiming at making AI go well.
Hi LuOMM,
The idea is to have a mix. In the same Brief page, there is already an Explainer section, which is supposed to be as neutral as possible, but there are also other sections (the Contentious Points section, Common Questions, Q&A, etc) where users' different opinions and takes can emerge.
For example, let's imagine a hypothetical Brief on the OpenAI/ Hugging Face incident. In the Q&A section of this Brief, creator and journalist users can post questions like "How bad is this, really"? and AI safety researchers and orgs can post answers. The answers, as you can imagine, can turn out to be more like takes, so not very neutral at all.
in short, the idea is to have the possibility of both learning about the facts in an unbiased way, but in parallel also have access to all of the different takes, disagreements, etc on a particular topic or event.
The video walkthrough (like in the article) shows more details on this.
Let me know if you have any more questions or suggestions :)
Thanks so much for writing about this. I also strongly believe that there should be a lot more effort going in AI safety communications.
I'd like to share a few things:
1. My own recent project (just launched, also with an EA Forum post): telltheworld.io. I'm looking for all the feedback I can get, so it'd be great to get in touch or get warm intros with anyone in your network who could find this useful.
2. Another new AI safety comms project: https://commonsignals.org/
3. The Effective Communications WhatsApp group. If you're not already on it, you might want to join - lots of AI safety comms people there, which I believe share your same concerns.
Hi Karen,
I'm working on a new comms project in this exact direction, aimed at bringing more creators and journalists to AI safety communications. I just posted about it a few days ago, here on the EA Forum. Would love to have your take on it, if it's of any interested (not sure I can post a direct link in here, as I'm relatively new to the Forum).
Hey, thanks for the link.
I know there are already projects automating this kind of summarisation work. However, that is not what I'm going for with Tell The World.
The goal with Tell The World is not only to make this kind of material more easily digestible, but also to have creators and journalists be able to confidently and credibly deliver it to their audiences. That is only possible if sources come with a "stamp of approval" by human experts - which is what this project provides.
Another major goal is to get more creators and journalists outside of the AI safety space to learn about & cover AI catstrophic risk, and - why not - become part of a larger community aiming at making AI go well.
That said, thanks for offering to help :)