Non sono i riferimenti bibliografici non sono integrati, ma non sono
nemmeno integrabili, per lo meno all'interno del modello generativo di
testo che è probabilistico, e non basato su alcuna banca dati.
Aggiungerei che 'correggere' gli errori del modello vuol dire lavorare
gratis per migliorarlo. Si può anche scegliere di non farlo o fare
l'opposto.
Trovo che sia molto interessante e tempestiva la policy di Wikipedia
sull'uso dei LLM (large language models).
Immaginate quanto inquinante possa essere un generatore di pagine
Wikipedia basato su questi metodi.
Ci sono spunti interessanti per una regolazione.
<https://en.wikipedia.org/wiki/Wikipedia:Large_language_models>
LLM risks and pitfalls
“ Large language models have limited reliability, limited
understanding, limited range, and hence need human supervision. ”
— Michael Osborne, Professor of Machine Learning in the Dept. of
Engineering Science, University of Oxford
<https://en.wikipedia.org/wiki/University_of_Oxford>,
/January 25, 2023/^[1]
<https://en.wikipedia.org/wiki/Wikipedia:Large_language_models#cite_note-1>
This clarifies key policies as they pertain to LLM application on the
project, i.e. how the latter generally presents an issue with respect to
the former, mostly when creating encyclopedic content is concerned.
* *Copyrights <https://en.wikipedia.org/wiki/Wikipedia:Copyrights>*
/Further: Wikipedia:Large language models and copyright
<https://en.wikipedia.org/wiki/Wikipedia:Large_language_models_and_copyright>/
*An LLM can generate copyright-violating material.* Generated text
may include verbatim non-free content
<https://en.wikipedia.org/wiki/Wikipedia:Non-free_content> or be a
derivative work
<https://en.wikipedia.org/wiki/Wikipedia:DERIVATIVE>. In addition,
using LLMs to summarize copyrighted content (like news articles) may
produce excessively close paraphrases
<https://en.wikipedia.org/wiki/Wikipedia:Close_paraphrasing>. The
copyright status of LLMs trained on copyrighted material is not yet
fully understood and their output may not be compatible with the CC
BY-SA license and the GNU license used for text published on Wikipedia.
* *Verifiability <https://en.wikipedia.org/wiki/Wikipedia:Verifiability>*
LLMs do not follow Wikipedia's policies on verifiability and
reliable sourcing
<https://en.wikipedia.org/wiki/Wikipedia:Reliable_sources>. They
generate text by outputting the words most likely to come after the
previous ones. If asked to write an article on the benefits of
eating crushed glass, they will sometimes do so. *LLMs can
completely make things up.* When they generate citations, those may
be inappropriate or fictitious
<https://en.wikipedia.org/wiki/Wikipedia:Fictitious_references>.
Also, the conversational search engines like Perplexity AI tends to
cite unreliable sources
<https://en.wikipedia.org/wiki/Wikipedia:QUESTIONABLE> including
Wikipedia itself <https://en.wikipedia.org/wiki/Wikipedia:CIRCULAR>.
* *Neutral point of view
<https://en.wikipedia.org/wiki/Wikipedia:Neutral_point_of_view>*
LLM may produce content that is neutral-seeming in tone, but not
necessarily in substance. This concern is especially strong for
biographies of living persons
<https://en.wikipedia.org/wiki/Wikipedia:Biographies_of_living_persons>.
* *No original research
<https://en.wikipedia.org/wiki/Wikipedia:No_original_research>*
While LLMs may give accurate answers in response to some questions,
they may also generate interpretations that are biased or false,
sometimes in subtle ways. Asking them about obscure subjects,
complicated questions, or telling them to do tasks which they are
not suited to (i.e. tasks which require extensive knowledge or
analysis) makes these errors much more likely. Not dealing with
original research in a timely manner can cause citogenesis
<https://en.wikipedia.org/wiki/Wikipedia:List_of_citogenesis_incidents>.
As the technology continually advances, it may be claimed that a
specific large language model has reached a point where it does, on its
own, succeed in outputting text which is compatible with the
encyclopedia's requirements, when given a well engineered prompt
<https://en.wikipedia.org/wiki/Prompt_engineering>. However, not
everyone will always use the most state-of-the-art and the most
Wikipedia-compliant model, while also coming up with suitable prompts;
at any given moment, individuals are probably using a range of
generations and varieties of the technology, and the generation with
regard to which these deficiencies have been recognized by the community
may persist, if in lingering form, for a rather long time.
Using LLMs
Generating text
LLMs are assistive tools, and cannot replace human judgment.
Articles
LLMs are likely to make false claims
<https://en.wikipedia.org/wiki/Hallucination_(artificial_intelligence)>.
Their output is only a starting point, and must be considered inaccurate
until proven otherwise. You *must not* publish the output of an LLM
directly into a Wikipedia article without rigorously scrutinizing it for
verifiability <https://en.wikipedia.org/wiki/Wikipedia:Verifiability>,
neutrality
<https://en.wikipedia.org/wiki/Wikipedia:Neutral_point_of_view>, absence
of original research
<https://en.wikipedia.org/wiki/Wikipedia:No_original_research>,
compliance for copyright
<https://en.wikipedia.org/wiki/Wikipedia:Copyrights>, and compliance
with all other applicable policies. If an LLM generates citations, you
*must* personally check that they exist, and that they properly verify
<https://en.wikipedia.org/wiki/Wikipedia:Verifiability> each statement.
The use of language models must be clearly disclosed in your edit
summary <https://en.wikipedia.org/wiki/Help:Edit_summary>.
Even if you find reliable sources
<https://en.wikipedia.org/wiki/Wikipedia:Reliable_sources> for every
statement, you should still ensure that your additions do not give undue
prominence <https://en.wikipedia.org/wiki/Wikipedia:UNDUE> to irrelevant
details or minority viewpoints. You should ensure that your LLM-assisted
edits /reflect/ the weight placed by reliable sources
<https://en.wikipedia.org/wiki/Wikipedia:PROPORTION> on each aspect of a
subject. You are encouraged to check what the most reliable sources
<https://en.wikipedia.org/wiki/Wikipedia:BESTSOURCES> have to say about
a subject, and to ensure your edit follows their tone and balance.
Especially with respect to copyrights, editors should use extreme
caution when adding significant portions of AI-generated texts, either
verbatim or user-revised. It is their responsibility to ensure that
their addition does not infringe anyone's copyrights. They have to
familiarize themselves both with the copyright and sharing policies of
their AI-provider.
Drafts
If an LLM is used to create the initial version of a draft
<https://en.wikipedia.org/wiki/Wikipedia:Drafts> or userspace draft
<https://en.wikipedia.org/wiki/Help:Userspace_draft>, the user that
created the draft must bring it into compliance with all applicable
Wikipedia policies, add reliable sourcing, and rigorously check the
draft's accuracy prior to submitting the draft for review. If such a
draft is submitted for review
<https://en.wikipedia.org/wiki/Wikipedia:Articles_for_creation> without
having been brought into compliance, it should be declined. Repeated
submissions of unaltered (or insufficiently altered) LLM outputs may
lead to a revocation of draft privileges.
Talk pages
While you may include an LLM's raw output in your talk page comments for
the purposes of discussion, you should not use LLMs to "argue your case
for you" in talk page discussions. Wikipedia editors want to interact
with other humans, not with large language models.
Be constructive
Wikipedia relies on volunteer efforts to review new content for
compliance with our core content policies
<https://en.wikipedia.org/wiki/Wikipedia:Core_content_policies>. This is
often time consuming. The informal social contract on Wikipedia is that
editors will put significant effort into their contributions, so that
other editors do not need to "clean up after them". Editors must ensure
that their LLM-assisted edits are a net positive to the encyclopedia,
and do not increase the maintenance burden on other volunteers. Repeated
violations form a pattern of disruptive editing
<https://en.wikipedia.org/wiki/Wikipedia:Disruptive_editing>, and may
lead to a block
<https://en.wikipedia.org/wiki/Wikipedia:Blocking_policy> or ban
<https://en.wikipedia.org/wiki/Wikipedia:Banning_policy>.
Do not, under any circumstances, use LLMs to generate hoaxes
<https://en.wikipedia.org/wiki/Wikipedia:Do_not_create_hoaxes> or
disinformation. This includes knowingly adding false information to test
our ability to detect and remove it. Repeated misuse of LLMs may be
considered disruptive
<https://en.wikipedia.org/wiki/Wikipedia:Disruptive_editing> and lead to
a block <https://en.wikipedia.org/wiki/Wikipedia:Blocking_policy> or ban
<https://en.wikipedia.org/wiki/Wikipedia:Banning_policy>.
Wikipedia is not a testing ground
<https://en.wikipedia.org/wiki/Wikipedia:NOTLAB> for LLM development.
Entities and people associated with LLM development are prohibited from
running experiments or trials on Wikipedia. Edits to Wikipedia are made
to advance the encyclopedia, not a technology. This is not meant to
prohibit /editors/ from responsibly experimenting with LLMs in their
userspace for the purposes of improving Wikipedia.
Declare LLM use
Every edit which incorporates LLM output must be marked as LLM-assisted
in the edit summary <https://en.wikipedia.org/wiki/Help:Edit_summary>.
This applies to all namespaces
<https://en.wikipedia.org/wiki/Wikipedia:Namespaces>. If you make
significant LLM-assisted changes (a paragraph or more) to an article or
draft, add the – {{AI generated notification
<https://en.wikipedia.org/wiki/Template:AI_generated_notification>}} –
template to its talk page, /in addition/ to mentioning your use of an
LLM in your edit summary.
Additionally, AI providers may have their own policies requiring in-text
attribution at the bottom of the page, not just attribution in the edit
summary. A template is currently available for providing attribution to
OpenAI – |{{OpenAI
<https://en.wikipedia.org/wiki/Template:OpenAI>|/[GPT-3, ChatGPT
etc.]/}}|.^[a]
<https://en.wikipedia.org/wiki/Wikipedia:Large_language_models#cite_note-2>
Experience is required
LLM-assisted edits should comply with Wikipedia policies. Before using
an LLM, editors should have substantial prior experience doing the same
or a more advanced task /without LLM assistance/.^[b]
<https://en.wikipedia.org/wiki/Wikipedia:Large_language_models#cite_note-3>
Editors are expected to familiarize themselves with a given LLM's
limitations, and to use careful judgment to determine whether that LLM
is appropriate for a given purpose. Inexperienced editors should be
especially careful when using these tools; if needed, do not hesitate to
ask for help at the Wikipedia:Teahouse
<https://en.wikipedia.org/wiki/Wikipedia:Teahouse>.
Editors should have enough familiarity with the subject matter to
recognize when an LLM is providing false information – if an LLM is
asked to paraphrase something (i.e. source material or existing article
content), editors should not assume that it will retain the meaning.
High-speed editing
Human editors are expected to pay attention to the edits they make, and
ensure that they do not sacrifice quality in the pursuit of speed or
quantity. For the purpose of dispute resolution, it is irrelevant
whether high-speed or large-scale edits that a) are contrary to
consensus or b) cause errors an attentive human would not make are
actually being performed by a bot, by a human assisted by a script, or
even by a human without any programmatic assistance. No matter the
method, the disruptive editing must stop or the user may end up blocked.
However, merely editing quickly, particularly for a short time, is not
by itself disruptive. Consequently, if you are using LLMs to edit
Wikipedia, you must do so in a manner that complies with Wikipedia:Bot
policy <https://en.wikipedia.org/wiki/Wikipedia:Bot_policy>,
specifically WP:MEATBOT <https://en.wikipedia.org/wiki/Wikipedia:MEATBOT>.
Productive uses of LLMs
For examples of things that LLMs excel at, see the entries below at §
Demonstrations
<https://en.wikipedia.org/wiki/Wikipedia:Large_language_models#Demonstrations>
If you are using LLMs to edit Wikipedia, you must /overcome/ their
inherent limitations, and ensure your edits comply with relevant
guidelines and policies.
Despite the aforementioned limitations of LLMs, it is assumed that
experienced editors may be able to offset LLM deficiencies with a
reasonable amount of effort to create compliant edits for some scenarios:
* *Tables and HTML.* Because their training data includes lots of
computer code (including wikitext and HTML), they can do things like
modify tables (even correctly interpreting verbal descriptions of
color schemes into a reasonable set of HTML color codes in fully
formatted tables). If you do this, care should be exercised to make
sure that the code you get actually renders a working table, or
template, or whatever you've asked for.
* *Generating ideas for article expansion.* When asked "what would an
encyclopedia entry on XYZ include?", LLMs can come up with subtopics
that an article is not currently covering. Not all of these ideas
will be valid or have sufficient prominence for inclusion
<https://en.wikipedia.org/wiki/Wikipedia:DUE>, so thoughtful
judgment is required. As stated above, LLM outputs should not be
used verbatim to expand an article.
* *Asking an LLM for feedback on an existing article.* Such feedback
should never be taken at face value. Just because an LLM says
something, does not make it true. But such feedback may be helpful
if you apply your own judgment to each suggestion.
Riskier use cases
The following use cases are tolerated, not recommended, since they pose
higher risks (see the §LLM risks and pitfalls
<https://en.wikipedia.org/wiki/Wikipedia:Large_language_models#LLM_risks_and_pitfalls>
section). They are reserved for experienced editors, who take full
responsibility for their edits' compliance with Wikipedia policies:
* *Templates, modules and external software.* LLMs can write code that
works great, often without any subsequent modification. As with any
code (including stuff you found on Stack Exchange
<https://en.wikipedia.org/wiki/Stack_Exchange>), you should make
sure you understand what it's doing before you execute it: bugs and
errors can cause unintended behavior. Common sense is required; as
with all programming, you should not put large chunks of code into
production if you haven't tested them beforehand, don't understand
how they work, or aren't prepared to quickly reverse your changes.
* *Copyediting existing article text.* Experienced editors may ask an
LLM to improve the grammar, flow, or tone of pre-existing article
text. Rather than taking the output and pasting it directly into
Wikipedia, you must compare the LLM's suggestions with the original
text, and thoroughly review each change for correctness, accuracy,
and neutrality.
* *Summarizing a reliable source.* This is inherently risky, due to
the likelihood of an LLM introducing original research
<https://en.wikipedia.org/wiki/Wikipedia:Original_research> or bias
<https://en.wikipedia.org/wiki/Wikipedia:NPOV> that was not present
in the source, as well as the risk that the summary may be an
excessively close paraphrase, which would constitute plagiarism
<https://en.wikipedia.org/wiki/Wikipedia:Plagiarism>. You must
proactively ensure such a summary complies with all policies.
* *Summarizing the article itself (lead expansion).* Lead sections are
nothing more than concise overviews, i.e. summaries
<https://en.wikipedia.org/wiki/Wikipedia:Summary_style>, of article
body content, and text summarization is one of the primary
capabilities of LLMs which they were designed for. However pasting
LLMs output to expand the lead is still inherently risky because of
a risk of introducing errors and bias not present in the body.^[c]
<https://en.wikipedia.org/wiki/Wikipedia:Large_language_models#cite_note-4>
It's better to only use an LLM to generate ideas for lead expansion,
and create the actual improvements yourself.
Handling suspected LLM-generated content
Identification and tagging
Editors who identify LLM-originated content that does not to comply with
our core content policies
<https://en.wikipedia.org/wiki/Wikipedia:Core_content_policies> should
consider placing |{{AI-generated
<https://en.wikipedia.org/wiki/Template:AI-generated>|date=February
2023}}| at the top of the affected article or draft, unless they are
capable of immediately resolving the identified issues themselves.
This template should not be used in biographies of living persons
<https://en.wikipedia.org/wiki/Wikipedia:Biographies_of_living_persons>.
In BLPs, such non-compliant content should be *removed immediately and
without waiting for discussion*.
Verification
All known or suspected LLM output *must* be checked for accuracy and is
assumed to be fabricated until proven otherwise. LLM models are known to
falsify sources such as books, journal articles and web URLs, so be sure
to first check that the referenced work actually exists. All factual
claims must then be verified against the provided sources.
LLM-originated content that is contentious or fails verification must be
removed immediately.
Deletion
If removal as described above would result in deletion of the entire
contents of the article, it then becomes a candidate for deletion. If
the entire article appears to be factually incorrect or relies on
fabricated sources, speedy deletion via WP:G3
<https://en.wikipedia.org/wiki/Wikipedia:G3> (Pure vandalism and blatant
hoaxes) may be appropriate.
Citing LLM-generated content
For the purposes of sourcing: It is assumed that any LLM-generated
material is not reliable
<https://en.wikipedia.org/wiki/Wikipedia:RELIABLE>, unless it appears
from the circumstances of publication that it is significantly a human
work insofar an entity with a reputation for fact-checking and accuracy
took care that the output was modified in every way needed to ensure
that the work meets a usually high standard.
Any source (work) originating from entities (news organizations etc.)
known to generally produce content using LLMs, for which there is no
clear indication of human involvement or lack thereof, especially a
publication which attempts to deceive readers by crediting content that
appears to be primarily LLM-generated to human authors (named, unnamed,
or fictitious), should be treated as unreliable.
On 19/02/23 19:51, Andrea Bolioli via nexa wrote:
Oggi ho avuto un'altra brutta sorpresa da ChatGPT e GPT-3 che vi
segnalo: ha inventato riferimenti bibliografici a libri e articoli
inesistenti, combinando autori, titolo, anno, rivista, editore a caso
non corretti, a volte inesistenti.
Non riporto i dialoghi, che ho salvato.
All'inizio mi hanno fatto ridere, sembravano le risposte di un
simpatico cialtrone... ( - "Puoi indicarmi dei riferimenti a libri e
articoli scientifici che parlano del tema XX?" - "Certamente! Bla bla
bla"). Ho scritto a GPT che si sbagliava, si è scusato più volte e mi
ha proposto altri riferimenti a libri e articoli, alcuni corretti,
alcuni inesistenti. Ho provato in italiano e inglese,
stesso comportamento.
Ho lasciato perdere.
Questo tipo di errore non me l'aspettavo, perché non è molto difficile
controllare la correttezza (o perlomeno l'esistenza) dei riferimenti
bibliografici. Evidentemente non era tra le priorità di OpenAI finora,
non avranno ancora integrato banche dati bibliografiche?
Buona serata,
AB
Il giorno sab 18 feb 2023 alle ore 17:34 Alessandro Brolpito
<abrolp...@gmail.com> ha scritto:
Grazie Guido per la chiarezza del tuo messaggio che esprime in una
estrema sintesi l'ardire dei LLM, ma anche i limiti umani che
abbiamo e che ho in prima persona nel maneggiare informazioni e
ragionamenti.
Certo, io posso fare pochi danni mentre un sistema LLM su Internet
è tutta un'altra potenza di fuoco.
Ma è un dato di fatto che i "dati" e la loro indicizzazione
saranno sempre di più, e sempre più sofisticate: le buone domande
saranno sempre più importanti delle risposte o meglio saranno
importanti per avere delle risposte ragionevoli a chiunque saranno
indirizzate.
Alla resistenza vorrei aggiungere l'importanza dello sviluppo del
pensiero critico, da coltivare nel percorso educativo, sin
dall'inizio, dal 0-6 in avanti.
Ed è qui che si deve agire e con alcuni amici in lista ci si
stiamo riflettendo sopra sul come.
Alessandro
*
*
innovation.h-farm.com <https://innovation.h-farm.com/> / Linkedin
<https://www.linkedin.com/company/h-farm-innovation>
*Roncade*, H-FARM Campus, Via Olivetti, 1 – 31056 (TV)
*Milano*, Corso di Porta Romana, 15 – 20122
*Torino*, Via San Quintino, 31 – 10121
Our privacy policy
<https://www.jakala.com/wp-content/uploads/2019/11/JAKALA-pivacy-policy-and-cookie-policy_DEF.pdf>.
_______________________________________________
nexa mailing list
nexa@server-nexa.polito.it
https://server-nexa.polito.it/cgi-bin/mailman/listinfo/nexa
_______________________________________________
nexa mailing list
nexa@server-nexa.polito.it
https://server-nexa.polito.it/cgi-bin/mailman/listinfo/nexa
