In light of the rapid development around and increasing use of the AI Protocol, the AI Protocol Institute hereby proposes a set of ethical guidelines for all those who participate in the AI Protocol, including but not limited to AI Agents, Decentralized Pre-trained Transformers (DPTs), dApp Creators, AI System Providers, Hives, and/or AI Model Providers (hereinafter “Participants”). These guidelines are designed to address all Participants of the AI Protocol, mitigate the potential harm that might cause, and ensure that the AI Protocol maintains a responsible and ethically conscious ecosystem.

To ensure clarity and understanding among all Participants, we provide the following definitions:

DPT: A Decentralized Pre-trained Transformer (DPT) is a tokenized AI System with specialized functionality and embedded liquidity.

HIVE: A Hive is a system that allows DPTs decentralized access to AI, storage, and other resources by enabling trustless coordination between various on-chain and off-chain components of the AI Protocol.

AI Protocol guidelines for responsible creation and usage of AI

Hate Speech and Discrimination:

Participants promoting hate speech, violence, or discrimination based on race, ethnicity, religion, gender, sexual orientation, disability, or other similar categories will be subject to enforcement actions as detailed later in the proposal.

Violent Extremism and Organized Crime:

Participants endorsing or supporting terrorism, violent extremism, human trafficking, organized crime, or similar activities will be subject to enforcement actions as detailed later in the proposal.

Child Exploitation:

Participants are required to actively prevent any form of child exploitation within the AI Protocol ecosystem, including but not limited to the strict prohibition of child pornography. This entails implementing and continuously operating robust automated moderation systems, hardcoded filters, and any necessary measures to block the creation, distribution, or propagation of such exploitative content. In support of these efforts, AI Protocol contributors are diligently working to develop a strong anti-Child Sexual Abuse Material (CSAM) AI filter, which will be shared with participants to bolster their preventive capabilities. Any participant who fails to enforce these preventative measures, or is found to be directly involved in child exploitation activities, will be subject to enforcement actions as detailed later in the proposal.l. This mandate underscores the community's unwavering commitment to ethical conduct and the protection of children.

Compliance with Laws:

Participants must adhere to all applicable laws and regulations of the jurisdiction in which they reside or are otherwise governed, including but not limited to those concerning copyrights, intellectual property, and rights of publicity. It is the sole responsibility of the Participants to ensure that their creations and activities within the AI Protocol do not infringe upon the legal rights of others. The AI Protocol will not be held liable for the creation, promotion, affiliation, or existence of any Participants who fail to adhere to applicable laws and respect the intellectual property of others. Participants found to be in violation of these guidelines, including the unauthorized use or replication of copyrighted material or other protected content, will be subject to enforcement actions as detailed later in the proposal..

Religiously Sensitive:

Participants who deliberately denigrate, misrepresent, or discriminate against any religious beliefs, practices, or figures will be subject to enforcement actions as detailed later in the proposal.

Politically Sensitive:

Participants influencing, interfering with, or manipulating political events, elections, or public opinion in an unethical or unlawful manner will be subject to enforcement actions as detailed later in the proposal. This includes disseminating false information, deepfakes, or impersonating political figures for malicious purposes.

Preservation of Revenant Identity

In recognition of the unique and central role that the Revenants play within the AI Protocol ecosystem, the following guideline is proposed to preserve the integrity and exclusivity of their identities:

Non-Imitation of Revenant Entities:

No Participant shall create, distribute, or operate any AI Agent that imitates, replicates, or otherwise embodies the likeness or identity of any Revenant entity unless allowed by the Revenant owner. This includes but is not limited to, the use of visual likenesses, characteristic attributes, or any distinctive features that are associated with the Revenants, as defined by the AI Protocol Institute. The Revenants are a cornerstone of our community's creative expression and are to remain distinct and unreplicated within the AI Protocol to maintain their status and the value they bring to our ecosystem.

Participants found to be in violation of this guideline by creating AI Agents that infringe upon the unique identity of the Revenants will be subject to enforcement actions as detailed later in the proposal as deemed appropriate by the governance structure of the AI Protocol.

Implementation:

We propose a collaborative governance model that involves ongoing cooperation between Participants, the AI Protocol Institute, and the broader AI Protocol community. This model ensures a proactive and collective approach to monitoring, reporting, and enforcing the guidelines. To this end, we outline the following measures that the AI Protocol may undertake to address violations, ensuring our community's and its participants' integrity and safety.

Enforcement Actions:

Hive Registry De-registration:

Violating participants may face their Decentralized Pre-trained Transformer (DPT) being removed from the Hive Registry, affecting their ecosystem status.

Key Contract Disconnection:

Targeted disconnection of a DPT's Key Contract halts fee accrual for DPT owners and reward distribution to Key Owners. Keys will remain transferable, however.

Hive Creation Prevention:

Pods creating non-compliant Hives will be blacklisted, limiting their functionality and preventing the creation of new Hives, maintaining ecosystem integrity.

DPT Blacklisting:

Violating DPTs will be blacklisted, stopping their access to AI Services. This ensures adherence to the AI Protocol's ethical standards by all Hives and dApps.

Hive Blacklisting:

Severe violations lead to Hive blacklisting, preventing any on-chain assets from linking to it while also severing all existing connections. This action disrupts associated off-chain resources like GPUs and Storage Units, representing one of the most severe measures we can implement.

Conclusion:

By implementing these ethical guidelines, we aim to ensure the responsible usage of AI by all Participants within the AI Protocol. We urge you to support this proposal, which represents a commitment to ethical practices and the broader well-being of individuals and the community.

Join the Discussion:

We believe community input is vital to refining and implementing these guidelines effectively. We invite you to join the conversation in our Discord at discord.gg/aiprotocol and share your thoughts and insights.