No More Dangerous Visions: AI Guardrails and Humanities Work

 

After eight years, this blog creaks, groans, stretches, and returns to life not about virtual worlds, but the newest technology on the block: Generative AI.

 Today my students put essays to two different Large Language Models (LLMs) hosted on our campus. All of them first used Claude Sonnet 3.5 and then the groups divided, with half putting the essays to ChatGPT 4 and the others Google Gemini Pro.  By the time you read this, the versions and even names may have changed.

In any case, I asked the students to provide the assignment prompt from a partner's paper, as well as the text of the paper itself. Then the students were to compare how the LLMs' advice compares to that of a human partner as well as Grammarly. 

For three students with Humanities topics, Anthropic's Claude refused, like Barletby the Scrivener in Melville's story, to do any work. It locked the students out of the software and logged them off.

Why? Subject matter and explicit language quite appropriate when quoting from some literary works or discussion topics such as human trafficking or sexual violence. So then we quizzed Claude about why. I asked it a general question about which words or phrases it flags as "unsafe" and got this reply:

As an AI language model, I don't actually have a system for identifying "unsafe content" or refusing to answer based on specific words or phrases. I'm designed to engage in a wide range of conversations while maintaining ethical boundaries. However, I do have guidelines that help me respond appropriately to different types of queries: 

1. I don't engage in or promote illegal activities. 
2. I avoid producing explicit sexual content or pornography. 
3. I don't assist in creating harmful content like hate speech or extreme political propaganda. 
4. I don't share personal information about real individuals. 
5. I avoid giving medical, legal, or financial advice that should come from professionals. 

To avoid any issues, it's generally best to: 

1. Keep queries legal and ethical. 
2. Avoid requests for explicit sexual content. 
3. Don't ask for help with harmful or dangerous activities.
4. Stick to general knowledge rather than specific personal information. 
5. For professional advice, consult experts in the relevant field. 

Remember, I'm here to help with a wide range of topics, and I'll do my best to provide useful information while staying within appropriate boundaries. If a query is problematic, I'll typically explain why I can't assist and suggest alternative approaches if possible.

One of my students whose partner's project got flagged as "unsafe" got the AI to provide further details for that one case:

I understand your frustration with the content being flagged. While I don't have access to the specific algorithms used to flag content, there are certain words and phrases commonly associated with explicit sexual content that might trigger such flags. These often include: 

• Anatomical terms for genitalia 
• Explicit verbs describing sexual acts 
• Slang terms for sexual activities 
• Graphic descriptions of bodily fluids or functions 
• Certain words associated with pornography or sex work 

To work around this, you could try: 

• Using more clinical or academic terminology 
• Focusing on the themes and concepts rather than specific acts 
• Summarizing the content instead of quoting directly 
• Using euphemisms or less explicit language 

If you'd like, you can try describing the general topic and main points of the paragraph without using potentially flagged words. I can then suggest how to phrase it in a way that's likely to pass content filters while still conveying the essential information.

Thus an AI would not be able to work with some of Allen Ginsberg's or Walt Whitman's poetry, Toni Morrison's Beloved, and many other literary works, let alone news accounts of many topics that contain graphic language or imagery. Such prudence appears well intentioned but bothersome to this writer, since our purpose was not to titillate a reader but to inform and educate with the help of the cybernetic co-pilot. 

We very much need Humanists, not just Business and STEM folks, using AI in canny ways. Some of our work will involve dangerous topics. We Humanists ask some very hard questions not germane in other fields of study.

Good. AI and its training data need cross-examination. I'm heartened that issues of sustainability and social justice have begun to be considered. All technologies include unforeseen social and economic consequences.

Before we part, back to why I revived "In a Strange Land." I feel the same Uncanny Valley moment I did in 2007, when I first logged into a virtual world. As with that, AI reveals new and sometimes unsettling terrain.

Incidentally, I won't rule out future posts about virtual worlds, though Iggy my Second Life avatar gave up his virtual land and wanders penniless around the once-hyped simulation. I also dip into a roleplaying game build inside SL occasionally, playing a virtual pirate, to conduct naval battles with others interested in The Age of Sail. 

Avast! The Royal Navy will never get me!

Images: Prompt to Dall-E, "Female undergraduate working in library with AI Copilot" and from Second Life, the Dread Pirate Rumbottom O'Malley and Gray Nick the parrot.

The Illusion Is Real To Me

This post began with some musings at Wagner James "Hamlet" Au's blog, New World Notes. Au has also covered the utopians and skeptics of virtual reality in a piece at Wired, "VR Will Make Life Better--Or Just Be an Opiate For the Masses." Some proponents of the Occulus Rift 3D viewer are claiming that a virtual world good enough is as good at the life we lead without goggles and a fast internet connection.

Since this issue involves both of my blogs, I hope readers will excuse the cross-postings. I have also been thinking about how Mark Zuckerberg of Facebook is very much invested in Occulus tech and why the picture of him, striding among a crowd wearing Rifts, chills me to the bone.

Here was my reply at New World Notes.

Even a happy virtual life would neglect the agons of a happy real one.

Yesterday I labored a few hours on a farm tractor I'm selling, checking for issues and eliminating possibilities in the electric system. I used a lot of stored knowledge in the wetware of my brain to trouble-shoot. I was in a place without reliable wireless, so double-checking hunches with the phone was not possible.

Then I went to work on the fuel system, turning wrenches and skinning knuckles until I had the likely culprit. At night I went online, into a flat virtual community, to check my assumptions. This weekend I'll clean out the fuel tank, blow compressed air through all the fittings, and restart the old diesel.

Simulating all that with an Occulus might eventually be possible. Doing so might even feed me if my virtual farm supplied RL income. But you know what?

Virtual is still FAKE. Always will be until someone really does achieve the Singularity. Hence my consideration of SL and more advanced forms of virtual worlds as just something nice for entertainment, like a novel or film but more immersive.

As for the possibility that our RL world is a Matrix? Let me quote a famous fake person, Conan the Barbarian. I only slay groundhogs and cold beers, but the rest is apt:

"I know this: if life is illusion, then I am no less an illusion, and being thus, the illusion is real to me. I live, I burn with life, I love, I slay, and am content."

Dutch Boy Running Out of Fingers at Linden Lab

Location: Back at VWER Roundtable

How odd it feels to be blogging again, ever so briefly, about Second Life. I have returned to the virtual world for a couple of VWER meetings and am even considering updating Iggy's avatar shape and skin. His relatively new dreadlocks demand it!

Yet following a post I spotted in New World Notes, about Linden Lab's deciding to lower set-up fees for sims, I thought my two cents might contribute to the debate.

Second Life continues to lose sims at a stately pace. I wonder, as many SLers do, if the entire world is no more than a cash-cow for the Lab to milk until Project Sansar launches. Purportedly, it will:

"democratize virtual reality as a creative medium. It will empower people to easily create, share, and monetize their own multi-user, interactive virtual experiences, without requiring engineering resources. The platform will enable professional-level quality and performance with exceptional visual fidelity, 3D audio, and physics simulation."

These promises are at odds. If Sansar lets us "easily" create such 3D content using our Occulus Rift headsets, it would require tools far simpler than Blender or Maya.  Those high-end tools then put Sansar out of reach of many educators I know.  Motion-sickness issues for the Rift may be easier to resolve than those about the tools needed for content creation. I reserve judgement on the "native building options" the Lab mentions. Perhaps in-world creation for ordinary mortals and student teams will endure, freed from the cumbersome permissions system that hamstrings SL team-builds.

Meanwhile, the Lab's original virtual world chugs along shedding 20-30 sims per week of landmass, rather like an iceberg drifting with the Gulf Stream. The Lindens do no, and probably can not, do the one thing that would democratize Second Life again: cut monthly tiers deeply for ordinary users.

 

Image of trends for private estates, from Gridsurvey.com

At $25 per month for a homestead and $50 for a full sim, I'd rent server space to make a roleplaying game from my twisted imagination. I could then take advantage of SL's rich marketplace, finely designed mesh avatars and other content. Linden Lab launched its Experience Keys program precisely for content like what I envision. Unlike 3D creations at Turbosquid, the SL marketplace offers content for pennies or just a few dollars. I could build the rest.

I've no faith that the Lindens would do what it takes to get tiers lower, such as moving to a low-rent neighborhood far from San Francisco's posh restaurants and boutiques. I've no faith they'd focus more staff time on finding out what ordinary SLers, and not just their land-barons, need.

It will be interesting to see what happens as Sansar launches. I fear that some SLers are so immersed in the world that they do not see the dyke cracking and the Dutch Boy trying every finger and toe he has to plug the leaks. Then, one day, comes the deluge.

VWBPE 2015: Staying the Course!

Location: VWER Meeting

I don't get Iggy into SL much any more, but I decided to pop in to hear about this year's VWBPE Conference, one I attended when I was more active in my use of virtual worlds.

This year's conference sessions look really interesting. Sadly, I'll miss it, since I'm going to be participating virtually, via Skype, in the CCCC 2015 Conference for writing teachers that week.

There's some irony there: me presenting via my RL self about work done a few years back in a virtual world. I would love to attend the SL conference, too, but I'd need a clone.

The resilience of VWBPE is timely. Just this week, I was at an academic meeting where an article was mentioned by a colleague. The source? Journal of Virtual Worlds Research.

It's an indication that these worlds are getting closer to that academic mainstream, for scholarship if not for widespread use.  Perhaps that will follow.