13 Comments
User's avatar
Shmuel Lome's avatar

I think an important aspect of this, is that language models are really just a 'cool' version of Wikipedia. It’s biggest differentiators include the ability to cross-reference information and an appealing interface that emulates human convo. The amazing responses we receive from ChatGPT derive from the amalgamation of human input and the guidance provided by programmers and systems. (same can be said about image models) Like you wrote - none of AI material is entirely new.

The actual concern that keeps being repeated throughout Lichtenstein’s podcast is that AI can propagate incorrect information while sounding authoritative. This seems to be the root cause of this ban. Interesting that our rabbonim and so called 'technology leaders' mask it as an avodah zara charge while addressing the same issue the goyish velt has been addressing since day one. Just Google - AI and bad/ problem/ issue etc. and there are numerous individuals and institutions covering it. It is a speedily growing technology problem, not a kefirah one.

Regardless if they signed because of of A, B, or C doesn't matter. At the end of the day, this ban is obviously a failure to some extent, the least being frustration at an unclear statement against widely used technology. This saddens me due to fueling more erosion of trust in our rabbonim.

Thanks for sharing this. Looking forward to seeing you open more conversations.

This article was written by someone who is not so impressed by AI:

https://www.newyorker.com/science/annals-of-artificial-intelligence/there-is-no-ai

Edit: Spacing

Expand full comment
Yosef Hirsh's avatar

My mind was recently changed as AI has been used to generate NSFW pictures.

Expand full comment
Shmuel Lome's avatar

There were off brand systems that have been doing this since day one.

Expand full comment
Yosef Hirsh's avatar

no longer off brand....google the taylor swift story.....10 systems came up.

Expand full comment
Shmuel Lome's avatar

Pardon my brevity. Just like pornographic sites do not represent the internet as a whole, this holds true here as well.

Expand full comment
User's avatar
Comment deleted
Jul 13, 2023
Comment deleted
Expand full comment
Shmuel Lome's avatar

And who more equipped than them?

Expand full comment
Yosef Hirsh's avatar

🤷‍♂️

Expand full comment
User's avatar
Comment deleted
Jul 13, 2023
Comment deleted
Expand full comment
Shmuel Lome's avatar

Yosef,

1. As someone who knows ChatGPT quite well, I have no clue what you are referring to when saying "educate" oneself more than Google...? Vehamavin Lo Yavin.

2. Additionally, language models by definition are filtered as this is how you get the info (not random mashups), it just is not personalized for the user yet. OpenAI is actively filtering results which keeps people from doing inappropriate things. Also, there is what Gaurd Your Eyes has; what he claims is special but really just a system that puts in inital prompts to "keep it kosher" before you say hello.

3. Of course there are harmful ways to use technology, especially when it advances so quick. No one denies this but the pitfalls are quite obvious from going on or letting your child go on a girlfriend chatbot - just as obvious as not going on pornographic sites (and Snapchat in the first place).

3. At this point in time, AI language models are not advanced enough for anyone to take seriously... more of a cool thing. As Lichtenstein said "are we really that dumb" and I think kids are referenced here as well. I would argue that I would love to leave my kids on ChatGPT and nothing else internet related

Expand full comment
Yosef Hirsh's avatar

1. SexEd including deviant practices.

2.agreed

3.agreed

4.dont agree...I would keep them off both. Do you really want your 10 year old to be able to ask Chat GPT specific questions about sex?

Expand full comment
Shmuel Lome's avatar

Fair, so I would agree with you regarding children at this point in time. This is only because OpenAI has decided to leave borderline inappropriate unfiltered.

Expand full comment
Yosef Hirsh's avatar

I have changed my mind on the issue.

See the recent Taylor Swift AI news story

Expand full comment
Shmuel Lome's avatar

This can be more lenghtly but put concise - people abuse technology like always.

Expand full comment
Yosef Hirsh's avatar

true but now is essentially the same as open browser.....dont give it to kids.

Expand full comment