Today we are taking a closer look at Steem and some issues that deserve more attention than they have received so far.
This investigation focuses on community moderation, blockchain governance, commercial AI detection services, individual responsibility, transparency and user protection.
Above all, there is one fundamental question:
WHO WATCHES THE WATCHDOGS?
It is a question that every community claiming to support decentralization should be willing to address.
Using blockchain technology does not automatically guarantee that every decision made within the surrounding ecosystem is transparent, fair or free from concentrated influence.
Steem is an interesting project because of its DPoS architecture, its economic incentives and the opportunities it offers users to participate in a shared digital ecosystem.
However, like any other project, it also faces organizational, technical and human challenges.
Our objective is not to decide in advance who is right or wrong. It is to examine available information, identify unanswered questions and encourage a discussion based on verifiable evidence.
Because analysing a project also means being willing to examine what may not be working.
STEEM: THE BLOCKCHAIN IS DECENTRALIZED, BUT IS EVERYTHING AROUND IT?
We should begin with an important distinction.
Steem is a blockchain operating through a Delegated Proof of Stake (DPoS) consensus model.
STEEM Power holders can participate in the selection of Witnesses, who are responsible for producing blocks and supporting the operation of the blockchain protocol.
Witnesses are not simply community representatives with symbolic positions. They have an important technical responsibility within the network.
The official documentation describes their role and the voting mechanisms involved.
Official documentation:
https://developers.steem.io/it/glossary/
Official Steem whitepaper:
https://steem.com/wp-content/uploads/2025/07/whitepaper20250711.pdf
However, we must distinguish the blockchain itself from the applications, websites and services through which people interact with it.
A community, a web interface, a curation service and a moderation system do not necessarily belong to the blockchain's consensus layer.
This distinction matters.
A decision made by a community moderator should not automatically be attributed to the blockchain protocol. Likewise, the actions of an administrator do not necessarily represent a decision made by Steem Witnesses.
So we need to ask:
WHO ESTABLISHES THE RULES, WHO ENFORCES THEM, AND WHO CHECKS WHETHER THEY ARE BEING APPLIED CORRECTLY?
If a moderator uses a commercial AI detection service, we should know whether this is an individual decision, an internal community procedure or an official recommendation from the platform.
These are three different situations with potentially different responsibilities.
We cannot continue using the word "platform" as a generic explanation for every service, decision or action.
We need names, procedures, documentation and clear answers.
AI DETECTION TOOLS: HOW RELIABLE ARE THEIR RESULTS?
One of the most sensitive issues concerns commercial services that claim to identify AI-generated writing.
Over recent years, numerous tools have appeared. Some are free, others require subscriptions, and many provide percentages or confidence indicators that appear extremely precise.
The problem begins when these results are used to accuse individual authors.
Here is the central question:
CAN A RESULT GENERATED BY A COMMERCIAL DETECTION SERVICE BE TREATED AS DEFINITIVE PROOF OF A TEXT'S ORIGIN?
The scientifically responsible answer is that we cannot automatically attribute certification capabilities to these tools when such capabilities have not been demonstrated.
A particularly relevant study was published in December 2023 in the International Journal for Educational Integrity.
Title: Testing of detection tools for AI-generated text.
https://link.springer.com/article/10.1007/s40979-023-00146-z
The researchers examined twelve publicly available tools and two commercial systems.
Their findings identified significant reliability problems, including incorrect classifications involving both AI-generated and human-written content.
The study also highlighted how text modification and machine translation can further affect detection performance.
This is not simply a personal opinion. It is a published scientific investigation.
Another study, published in 2024 in Current Medical Research and Opinion, compared several AI detection services and found significant differences between the results produced by the tools examined.
https://pubmed.ncbi.nlm.nih.gov/38265047/
These findings explain why a percentage generated by a detector should never automatically become a final judgment about an author.
A detection service may provide an estimate based on its own model. That does not mean it has reconstructed the complete history of a document, identified its author or established the exact process used to produce it.
Those are very different things.
The problem with percentages
Imagine that a moderator submits the same article to three different services.
The first reports a high probability of AI-generated content.
The second returns an intermediate result.
The third considers the text predominantly human-written.
Which result should be accepted?
And what scientific criteria would justify choosing one service over another?
Purchasing a premium subscription does not transform a probabilistic result into a scientific or judicial certification.
A paid subscription may provide additional features, larger usage limits or more detailed reports. It does not automatically guarantee absolute accuracy.
The issue is not whether AI detection tools should exist.
The issue is whether support tools should be treated as authorities capable of conclusively establishing the authorship of a document.
What should a responsible verification process include?
Before making an accusation, moderators should consider:
The availability of the original draft.
Previous versions of the document.
Editing history.
The sources used.
Translation methods.
Editorial revisions.
Communication with the author.
The documented limitations of the detection service.
A detector may serve as an initial screening tool, but it should not replace a serious and documented investigation.
TERMS OF SERVICE, PRIVACY AND COPYRIGHT: WHAT USERS SHOULD KNOW
There is another issue that deserves attention.
Using a commercial service means more than simply entering text into an online box.
Users may be accepting contractual terms, privacy policies and provisions governing how uploaded content is handled.
These conditions vary from service to service.
Some providers may impose restrictions concerning document retention, data processing, third-party content or the rights associated with uploaded material.
We should not assume that all services have identical conditions, nor should we claim that every upload automatically constitutes a legal violation.
The actual terms of the specific service must be examined.
Anyone using these tools to analyse another person's content should first consider:
What information is being uploaded.
Whether the service stores submitted documents.
The purposes for which the data may be processed.
Whether restrictions apply to third-party material.
Which responsibilities are assigned to the user.
What guarantees the provider actually offers.
Copyright is another important consideration.
Publicly accessible content is not automatically free to use for every possible purpose.
At the same time, analysing another person's text does not automatically constitute copyright infringement.
The answer depends on the content, the circumstances of use, the applicable contractual conditions and the relevant law.
The principle is straightforward: understand the service before using it.
And understand the limitations of a result before publicly accusing someone on the basis of that result.
WHEN AN AUTOMATED RESULT BECOMES A PUBLIC ACCUSATION
There is a substantial difference between using a service for preliminary screening and publicly claiming that a particular author has used artificial intelligence.
The distinction matters because the second situation may affect someone's reputation.
We should distinguish between:
A suspicion.
An automated result.
A personal assessment.
A documented fact.
A scientifically supportable conclusion.
A finding established through the appropriate legal procedures.
These are not interchangeable.
In Italy, Article 595 of the Criminal Code governs defamation. Its third paragraph addresses offences committed through the press or other means of publicity.
The Italian Senate has also published an explanation of online defamation and the application of the relevant legal framework to communications through the internet.
Reference: https://www.senato.it/show-doc?id=1176495
Official legislation: https://www.normattiva.it/eli/id/1930/10/26/030U1398
This does not mean that every disagreement about an article constitutes defamation.
The specific wording, context, publication, alleged conduct and other legal requirements must be assessed.
However, it does mean that people should carefully consider the consequences of publishing allegations about identifiable individuals.
AN AUTOMATED SCORE SHOULD NOT BE PRESENTED AS CONCLUSIVE PROOF WITHOUT THE NECESSARY EVIDENCE.
This principle applies equally to moderators, curators, administrators and ordinary users.
RELIGIOUS DISCRIMINATION: WHERE DOES FREEDOM OF EXPRESSION END?
Another reported incident concerns content published within a Steem community.
The discussion identified by the author is available here:
https://steemit.com/hive-154900/@wakeupkitty/tnsgz1
The complaint concerns statements relating to Catholicism and its religious representatives.
According to the interpretation presented by the person raising the issue, the content may have conveyed misleading information about Catholic religious leaders and their alleged teachings, without providing sufficient supporting evidence.
The central question should not be which religion is right, which is wrong or which personal beliefs people should adopt.
Freedom of expression includes the ability to criticise religions, institutions, doctrines and religious representatives.
However, legitimate criticism and propaganda or incitement to religious discrimination are not the same thing.
In Italy, Article 604-bis of the Criminal Code addresses specific forms of propaganda and incitement to discrimination or violence on racial, ethnic, national or religious grounds.
Official reference:
https://www.giustizia.it/giustizia/it/mg_1_2_1.page?contentId=SAN49542
It is important to clarify that criticism of a religion does not automatically constitute this offence.
The legal assessment depends on the actual content, its context and the elements required by the relevant provision.
This is precisely why communities should approach such reports carefully.
If someone presents a statement as an established fact, they should be able to identify the evidence supporting it.
If another user challenges that interpretation, they should be able to do so without automatically being treated as a troublemaker.
And when a complaint is submitted to moderators, it should be assessed according to the applicable community rules.
Transparency must apply to everyone.
The responsibility of moderators
The question is not whether moderators should share a particular religious belief or personal opinion.
The question is whether community rules are applied consistently.
When potentially discriminatory content is reported, what checks are performed?
Is there a documented procedure?
Are users given an opportunity to respond?
Are moderation decisions explained?
Are the rules publicly available and consistently enforced?
These are legitimate questions, not personal attacks.
VOTES THAT ARE CAST AND LATER REMOVED: WHAT CAN WE ACTUALLY ESTABLISH?
Another reported incident concerns voting activity.
The material provided describes a case in which certain curators allegedly voted for a post and subsequently withdrew their votes.
The discussion identified is available here:
This behaviour deserves examination, but we must distinguish observable activity from motivations that cannot be established without further evidence.
A vote can be cast and subsequently withdrawn.
Blockchain records can help establish the sequence of transactions, but they do not necessarily reveal the personal reasons behind a decision.
A curator may change their mind, correct an assessment or decide to withdraw their support.
That action, by itself, does not prove misconduct.
The concern arises when there are allegations of coordinated behaviour, unequal treatment or decisions influenced by personal relationships.
In those circumstances, transaction histories, account comparisons, relevant operations and explanations from the people involved become important.
A sequence of events alone is not sufficient to prove coordination between multiple individuals.
But systematically ignoring reports of potentially unusual behaviour is not an appropriate response either.
TRANSPARENCY DOES NOT MEAN DENYING PROBLEMS. IT MEANS ALLOWING THEM TO BE VERIFIED.
WITNESSES, MODERATORS, CURATORS AND ADMINS: WHO ACTUALLY HAS AUTHORITY?
This may be the most important part of the entire investigation.
A DPoS ecosystem contains different roles with different responsibilities.
Witnesses participate in blockchain consensus and block production.
Administrators manage the organizational structures or services for which they are responsible.
Moderators enforce community rules.
Curators evaluate and vote on content according to their own decisions and any applicable community guidelines.
These roles must not be confused.
A community moderator does not automatically become a Witness.
A curator does not automatically acquire governance authority over the blockchain.
A community administrator does not necessarily represent the entire Steem organization.
This distinction is essential when assigning responsibility for disputed decisions.
However, further questions remain.
How are these roles assigned?
What qualifications are expected?
Are there review procedures?
How are conflicts of interest handled?
Who can report misconduct?
And most importantly, who reviews the actions of those responsible for monitoring everyone else?
An academic study published in 2021 examined Witness activity within the Steem blockchain, highlighting their social importance and influence within the ecosystem.
Research: https://link.springer.com/article/10.1007/s11036-021-01749-x
Governance is not merely a technical matter.
It is also about trust, participation and accountability.
MULTIPLE ACCOUNTS, POWER CONCENTRATION AND VOTING TRANSPARENCY
Another issue worth examining concerns multiple accounts.
Having more than one account is not automatically a violation of blockchain rules.
However, legitimate multi-account use must be distinguished from possible coordinated activity intended to artificially influence voting, rewards or perceptions of community support.
This distinction is fundamental.
Identifying unusual behaviour requires more than observing that several accounts vote for the same posts.
Relevant evidence may include:
Transaction histories.
Delegation relationships.
Account connections.
Voting distributions.
Repeated patterns of behaviour.
Relevant economic relationships.
The resulting impact on rewards.
Even then, correlation does not automatically prove unlawful coordination.
Nevertheless, public blockchain data provides important opportunities for more detailed analysis.
This discussion is not limited to Steem.
We have previously examined similar governance and economic questions in Hive and Blurt, including concentration of influence, technical problems and the participation of ordinary users.
The objective is not to claim that every project has identical problems.
It is to recognize that DPoS systems should also be evaluated through their actual results.
WHO WATCHES THE WATCHDOGS? TEN QUESTIONS STEEM SHOULD ADDRESS
We have reached the central issue.
A project cannot rely solely on the statement that it is decentralized.
It must demonstrate that its participation and oversight mechanisms operate transparently.
For this reason, Steem should be willing to address several important questions.
Are AI detection services officially recommended?
If so, by whom and under which guidelines?Are there procedures for challenging automated results?
Can authors request human review and provide supporting documentation?Are moderators aware of the scientific limitations of AI detectors?
Are there guidelines preventing automated percentages from being treated as definitive proof?Which tools are actually being used?
Can their names, versions and applicable terms be disclosed?How are reports of religious discrimination handled?
Is there a consistent and documented procedure?How are votes that are subsequently withdrawn handled?
Can users obtain explanations when disputed voting behaviour raises legitimate questions?What safeguards exist against conflicts of interest?
Who checks that personal relationships do not influence moderation decisions?How are moderators and curators selected?
Are minimum qualifications, training and review procedures required?Who can intervene when a moderator is reported?
Is there an independent review mechanism separate from ordinary community management?What responsibilities should Witnesses and platform representatives have in moderation disputes?
The distinction between blockchain protocol, applications and individual community management must be clear.
These are not provocative questions.
They are questions that any project concerned about its credibility should be prepared to answer.
WHAT WE ARE ASKING FROM THE STEEM TEAM
We are not asking for privileges.
We are not asking for votes.
We are not asking for preferential treatment.
We are asking for transparency.
In particular, we believe the following points deserve clarification:
Whether official recommendations concerning AI detection services exist.
Which tools, if any, are recommended.
What scientific and operational limitations are acknowledged.
Whether users can challenge automated assessments.
How reports involving discrimination and defamation are handled.
What oversight mechanisms apply to moderators.
How conflicts of interest are addressed.
What procedures are available to improve transparency in community decisions.
We are not asking the team to intervene simply to support one person's position.
We are asking for clear procedures and consistent application of the rules.
If a tool is considered appropriate for a particular purpose, the reasons should be explained.
If it cannot provide definitive proof, its results should not be presented as a final certification.
The same standards should apply to every report involving published content and user behaviour.
STEEM, HIVE AND BLURT: DPoS PROJECTS SHOULD LEARN FROM PREVIOUS EXPERIENCES
Anyone who has followed these ecosystems for a long time knows that governance problems are not new.
Steem, Hive and Blurt have different histories, but they share certain technical and organizational characteristics that make comparisons interesting.
Decentralization does not automatically eliminate the concentration of influence.
Stake-weighted voting can create substantial differences in users' ability to influence decisions.
Communities can develop internal dynamics that are difficult to oversee.
Reward systems can encourage behaviours that differ from the original intentions of their designers.
And when trust disappears, even a technically functional platform can lose participation.
We have already explored these subjects in previous investigations
Related articles
STEEM: SOMETHING MUST CHANGE – 10 QUESTIONS FOR THE COMMUNITY ABOUT THE PLATFORM'S FUTURE
https://www.commentalanotizia.com/2026/10/steem-10-domande-futuro-community-dpos.html
HIVE: THE MILLION-DOLLAR PLATFORM UNDER SCRUTINY – MILLIONAIRE WITNESSES AND GHOST ACCOUNTS
https://www.commentalanotizia.com/2026/03/hive-piattaforma-milionaria-witness-fantasma.html
SHOCKING REVELATIONS ABOUT HIVE: MULTIPLE ACCOUNTS AND MASS DOWNVOTING
https://www.commentalanotizia.com/2026/03/rivelazioni-shock-su-hive-multi-account.html
BLURT: COLLAPSE OR TRANSFORMATION? BUGS, RISING COSTS AND NEW SCENARIOS
https://www.commentalanotizia.com/2026/09/blurt-collasso-trasformazione-bug-costi-nuovi-scenari.html
Each project must be assessed according to its own characteristics and verifiable evidence.
It would be incorrect to automatically transfer the responsibility for problems in other ecosystems to Steem.
But it would be equally wrong to ignore previous experiences.
OUR POSITION: NOBODY SHOULD BE ABOVE ACCOUNTABILITY
We have reached our conclusions.
This investigation is based on reported incidents, personal observations and questions that deserve answers.
We do not intend to present unverified allegations as established facts.
We do not intend to assign criminal responsibility to individuals or groups.
We do not intend to confuse community moderation with blockchain governance.
But we also do not intend to abandon our right to ask questions.
If an author is accused on the basis of an automated detector, they should be able to understand the evidence used and challenge the conclusion.
If a possible case of religious discrimination is reported, the community should examine it seriously.
If a vote is withdrawn, it is legitimate to ask for clarification when the circumstances raise questions.
If relationships between moderators, curators and administrators may influence decisions, it is legitimate to ask what safeguards exist.
And if someone believes a criticism is unfounded, they have every right to demonstrate that publicly through documents, data and explanations.
That is the kind of discussion we want.
Not a system in which everyone must agree.
Not a system in which people who ask questions are automatically treated as a problem.
Not a system in which decentralization becomes an excuse for avoiding accountability.
DECENTRALIZATION DOES NOT MEAN THE ABSENCE OF RESPONSIBILITY.
It also means distributing power, making decisions verifiable and allowing users to participate without fearing arbitrary treatment.
Steem still has an opportunity to demonstrate that its community can address these problems maturely.
But that requires answers, not merely statements of principle.
We need procedures, documentation, verification and, above all, a willingness to engage in open discussion.
Because everyone holding a position of responsibility should remember one thing:
NOBODY IS UNTOUCHABLE, AND NOBODY SHOULD BE ABOVE THE RULES.
The next step belongs to the community and, above all, to those responsible for managing the project.
We are not asking everyone to agree with our conclusions.
We are asking people to read, verify and respond to the questions raised.
A community grows when difficult questions are allowed to be discussed.
And when people in positions of responsibility demonstrate that they are not afraid of transparency.
Decentralization is built every day through concrete actions.
WHO WATCHES THE WATCHDOGS?
It is time to address this question seriously.
Thank you for reading.
🔗 ARTICLES WORTH READING
For further information about DPoS ecosystems and our previous investigations, we invite readers to explore the articles listed above.
SOURCES AND DOCUMENTATION
Steem official whitepaper:
https://steem.com/wp-content/uploads/2025/07/whitepaper20250711.pdf
Steem Developer Documentation:
https://developers.steem.io/
Scientific research on AI detection reliability: https://link.springer.com/article/10.1007/s40979-023-00146-z
Research comparing AI detection services: https://pubmed.ncbi.nlm.nih.gov/38265047/
Academic research on Steem Witnesses: https://link.springer.com/article/10.1007/s11036-021-01749-x
Italian Criminal Code:
https://www.normattiva.it/eli/id/1930/10/26/030U1398
Italian Ministry of Justice:
https://www.giustizia.it/giustizia/it/mg_1_2_1.page?contentId=SAN49542
Italian Senate – Online defamation:
https://www.senato.it/show-doc?id=1176495
FOLLOW OUR INVESTIGATIONS
Official website:
https://www.commentalanotizia.com/
Steemit publication archive:
https://www.steemit.com/@condividisulweb/posts
We will continue monitoring developments across Steem and other Web3 ecosystems, examining available evidence and encouraging open discussion.
👉 Want to read the full original article? Read it on Commenta La Notizia:👇
https://www.commentalanotizia.com/2026/11/steem-rilevatori-ia-moderazione-trasparenza-controllori.html
The original article is in Italian. You can use your browser’s automatic translation feature if needed.
Subscribe and vote for all articles 👇
https://www.publish0x.com/goldweb?a=xkazKWYYbJ
💬 Join us on Telegram to comment and follow all the insights
👉 https://t.me/CommentaLaNotizia
© 2026 CommentalaNotizia.com - All rights reserved
Official sources used under fair use
📺 Follow Commenta La Notizia for deeper insights, global analysis, and fearless discussions about the issues that shape our world — from politics and technology to society and the hidden forces behind the headlines.
⚠️ Legal Notice This article and the related multimedia content are the exclusive property of the author. Copying, reproduction, distribution, or modification in any form or on any platform is prohibited without written permission.
Copyright © CondividiSulWeb – YouTube Commenta La Notizia
👤 Author: @condividisulweb
✍️ Commenta La Notizia is my editorial project created to provide clear and engaging insights into news, events, and major current issues.
Every article, video, and short is made with a clear goal: not only to inform, but also to give context, reflection points, and analysis to understand not just what happens, but especially why it happens.
With Commenta La Notizia I want to build a reliable and accessible space where everyone can get informed, understand, and join the discussion.
🌐 If you want to stay updated, you can follow me here.
💬 Your opinion matters! Every piece of news is a chance to discuss: share your thoughts and make your voice heard.
📰 Comment, share, stay informed.
🙏 Thank you for reading: your time and your opinion are the true value
https://www.commentalanotizia.com/
