Re: Proposed draft of an AI policy

> I don't think it's worth possibly improving my own workflow by some small
amount
> by using a technology created using the underpaid labour of African
workers
> who are so traumatized by the work that they have symptoms of PTSD
(amongst many other ethical problems).
> That doesn't make me a monk or a Luddite.

Completely agreed, Bethan.

This is my actual response to a recruiter who is asking PhDs/Experts to
train AI models

==================================================================
> Please let me know your rate expectations on W2 on hourly basis?

Without knowing the exact workload and tasks it is pure gambling to name a
price ...

Say $100/hr. on W2?  :)

This reminds me of something that I definitely don't want to be part of -
therefore, could you, please, state that the proposed contract does not
fall into the gruesome scenarios cited below?

"Reports about the human labor used in training and moderating commercial
AI systems — especially content labeling and toxicity filtering outsourced
to low-paid workers in countries such as Kenya and other African nations.

The issue became widely known after investigations into companies
contracted to filter and annotate training data for major AI systems.

One of the most cited investigations involved workers in Kenya, Nigeria and
India, employed through outsourcing firms that provided data-labeling
services for OpenAI and others. Reports described workers reviewing:


   - graphic violence,
   - sexual abuse,
   - hate speech,
   - and other disturbing material,


often for relatively low wages and with limited psychological support.

This became a major ethical controversy because:


   - the work could reportedly cause trauma symptoms,
   - workers were economically vulnerable,
   - and the AI industry’s rapid growth depended partly on such hidden labor
   "

Thanks,
Dimitre Novatchev.
==============================================================

Maybe we can see clearly now that the ethical problem is exploitation and
harmful labor practices — not the mere existence of AI tools themselves.

Again, thanks for the interesting discussion,

Dimitre.

On Thu, May 28, 2026 at 10:02 AM Bethan Tovey-Walsh <
bytheway@linguacelta.com> wrote:

> Just for clarity, let's note that most of the examples you link aren't
> using LLMs; they're using machine learning models designed for maths. That
> isn't the same thing. Yes, some of the Erdős problems have been solved with
> help from ChatGPT. These are not particularly complex or difficult
> problems, as maths problems go, but it's still impressive.
>
> Regardless, the draft policy proposed by Norm doesn't stop individuals
> from using LLMs during the process of working out ideas and theories, if
> they feel their personal ethical code permits them to do so.
>
> So if you can overlook the ethical problems with LLMs, and you think
> you're somehow helped by using LLMs, then nothing in the draft policy stops
> you from developing your ideas with an LLM. It just asks that the actual
> submissions you make should not be produced by the LLM, for the reasons
> Norm outlined.
>
> Also, please stop the ad hominem attacks and the overblown language. My
> ethical principles mean I don't think it's worth possibly improving my own
> workflow by some small amount by using a technology created using the
> underpaid labour of African workers who are so traumatized by the work that
> they have symptoms of PTSD (amongst many other ethical problems). That
> doesn't make me a monk or a Luddite.
>
> As an autistic person, I supposedly lack empathy. Maybe that's right,
> because I cannot for the life of me empathize with colleagues who can play
> with technological toys while disregarding the pain of the people who are
> exploited to produce them, and who call me belittling names because my
> ethical principles won't permit me to ignore those exploited people.
>
> BTW
>
> ****************************************************
>
> Dr. Bethan Tovey-Walsh
>
> linguacelta.com
>
> Golygydd | Editor geirfan.cymru
>
> Croeso i chi ysgrifennu ataf yn y Gymraeg
>
> On 28 May 2026, at 17:26, Dimitre Novatchev <dnovatchev@gmail.com> wrote:
>
> 
> >  What LLMs are undeniably exceptionally good at is producing plausible
> text: text that a human reader will not immediately and obviously find
> flawed.
>
> Hmm...
>
> What about the Nobel Prize in Chemistry for the development of an AI model
> to solve the 50-year-old problem of protein folding?
> https://www.nobelprize.org/prizes/chemistry/2024/press-release/
>
> And the recent solution to a 60-year-old  Erdős'  problem - something that
> no mathematician could solve in 60 years?
>
> https://www.scientificamerican.com/article/amateur-armed-with-chatgpt-vibe-maths-a-60-year-old-problem/
>
> Not only gold medals at the International Mathematics Olympiad, but a
> useful accelerating tool in the work of many mathematicians, helping solve
> hard problems that have remained unsolved for years.
>
> https://www.quantamagazine.org/the-ai-revolution-in-math-has-arrived-20260413/
>
> Having all these results and the current process that produced them, the
> question arises:
> Is it wise for a group of people  to intentionally isolate themselves from
> a powerful tool that makes other people way more productive and helps them
> produce new results and solve some of the hardest problems in a
> record-short time? Form a closed society like monks in an isolated, remote
> monastery?
>
> Will it not be better to recognize the usefulness of these tools and start
> applying them in a controlled manner to achieve results better and faster,
> and even get unexpected ideas for new work?
>
> Like the church that ignored Galileo's tool (the telescope) results and
> had to admit its ignorance "only" 300+ years later?
> https://en.wikipedia.org/wiki/Galileo_affair
>
> Let us think before we act...
>
> Let us agree that the relevant distinction is not whether AI assisted in
> producing the text, but whether the contributor deeply understands and
> takes responsibility for the proposal.
>
> To summarise:
> The real unsolved question is whether or not a community can utilize the
> acceleration benefits while preserving accountability and semantic
> integrity?
>
>
> Dimitre.
>
> On Thu, May 28, 2026 at 4:25 AM Norm Tovey-Walsh <norm@saxonica.com>
> wrote:
>
>> Dimitre Novatchev <dnovatchev@gmail.com> writes:
>> > The first question that naturally arises is how to implement such a
>> policy in practice. Who, and using what mechanism, will judge
>> > the originality of the work?
>>
>> There’s only one practical way for the community group to implement any
>> policy: by what consensus the community group achieves.
>>
>> > This is a large and very important problem and at present entire
>> organizations and specialized events are focusing on it.
>>
>> It is a large and important problem and I’m glad that organizations are
>> working on it. There are lots of different aspects of the issue that we
>> could discuss.
>>
>> I would rather not. I think there *are* a host of very good moral,
>> ethical, environmental, and financial reasons to support the position that
>> commercial LLMs are a blight on our industry and on the world generally.
>> The first draft I wrote of the policy included a necessarily incomplete
>> enumeration of some of them. I took at that out. I don’t think we need to
>> debate those issues.
>>
>> I think there is a fundamental, technical issue that wholly and
>> decisively justifies a policy against contributions that originated in any
>> form of LLM.
>>
>> The task of our community group is to write a set of clear, coherent
>> specifications. This is an objectively difficult task. Human authors start
>> with a conceptual model of some feature that they believe the community
>> group agrees with, or that they wish to persuade the community group to
>> agree with. They attempt to describe that feature clearly in prose. Authors
>> are more-or-less successful at this for a wide variety of reasons some
>> within and some beyond their control.
>>
>> A reader starts with the prose and constructs their own model of the
>> feature. Like authors, readers are more-or-less successful at this for a
>> wide variety of reasons, the most significant of which is how clear and
>> understandable they find the author’s prose.
>>
>> Members of the community group can compare the model they constructed
>> from the prose with what they believe the community group discussed and/or
>> agreed. The small subset of readers who are implementing the specification
>> attempt to build software that performs the feature.
>>
>> The much broader community of readers decide if they believe they
>> understand the feature and if they think it is applicable to the use case
>> they have in mind. Then they compare what an implementation actually does
>> with what they thought it was going to do.
>>
>> The community compares notes. Authors and readers collaborate on building
>> an understandable specification. Sometimes defects are found in the
>> feature. Sometimes defects are found in the description of the model.
>> Sometimes they are found in implementations.
>>
>> The one through line in this entire process is that human authors are
>> attempting to communicate their understanding of a feature to human readers.
>>
>> Specification prose that I draft sometimes contains Americanisms that
>> some may find unfamiliar. Mike’s prose may sometimes contain Britishisms
>> that I find unfamiliar. An author writing in their second, third, or fourth
>> language may choose forms of expression that a native Enlish speaker would
>> not.
>>
>> Careful human review of all these drafts, despite their various
>> idiosyncracies, is valuable and justified by the fact that we are all
>> attempting, in good faith, to communicate technical ideas that we believe
>> we understand clearly, so that others can understand them.
>>
>> What LLMs are undeniably exceptionally good at is producing plausible
>> text: text that a human reader will not immediately and obviously find
>> flawed. But the LLM does not in any sense whatsoever *understand* the
>> features of our language.
>>
>> Asking humans to review text generated in this way is unduly burdensome.
>> Prose that reads well and appears to describe a feature, but on much closer
>> inspection is internally inconsistent, or worse, simply vacuous, is
>> corrosive to the process.
>>
>> Every one of us has a limited amount of time to read and review
>> specifications. We must use that time wisely and constructively. Reviewing
>> plausible text generated by a statistical process devoid of any actual
>> understanding of the specifications is, objectively, technically, a poor
>> use of our valuable time.
>>
>> We must have a policy that protects us from it.
>>
>>                                         Be seeing you,
>>                                           norm
>>
>> --
>> Norm Tovey-Walsh
>> CEO, Saxonica
>>
>>
>