Week 10

Mon 3rd Mar - Sun 9th Mar 2025

Monday 3rd March 2025

AI Ethics WG

Narrative:

  1. Interview Issues

Vani has checked most of the interviews for validity. Some have been sent for payment; others have some issues e.g. incorrect filenames, too short, no consent forms, etc. See individual interview Issues on the GitHub board for details. The rest will be checked by the end of the Quarter.

Payments will only be made if interviews meet the agreed validity criteria. We note that a short interview could still be included in the research corpus, provided there is a consent form for it - but in terms of WG payment, we have agreed that interviews should be 30 mins, so ones that are significantly shorter will be paid less. A sliding scale has been devised by the Admin team for interviews that are too short. See https://github.com/SingularityNet-Ambassador-Program/AI-Ethics-Workgroup/issues/117

We noted that regardless of interview validity, any expenses incurred by interviewees will still be paid.

What we've learned on interviewing:

  • We have noticed some difficulties for people working in a 2nd language - BGI Nexus will consider whether it is possible in future for interviews to be in languages other than English.

  • It is helpful to make a habit (as Clement did in his interviews) of avoiding any personally-identifiable info on the recording - e.g. not using the interviewee's name after the initial intro, etc. It makes anonymising easier if it's only the start of the interview that needs to be removed.

  • We need a more uniform approach to explaining who the organisation is that's doing the interviews - the nested "sNET/BGI Nexus/ Ambassadors/ AI Ethics WG" can be a bit of a mouthful! Vani will add a suggestion for what to say to the training slides and Interviewers' Handbook.

  • For group interviews, a consent form is needed for each person. Use the same identifier code, but with "a" "b" "c" on the end.

  • Selection of interviewees - we have found that it's best to choose people who have something to say. interviews with people who know nothing at all about AI have been difficult. But for future interviews, we could give some thought (perhaps in the flothcoming interviewers' plenary session) to how best to approach interviewees with no or very little knowledge of AI, so we can include them in our research.

  1. AI Ethics Discord Forum We agreed that since the forum isn't being very widely used, we will suspend paying for moderation of it next Quarter. The overal Discord moderation will pick up any problem posts.

  2. WG Sync Call We will mention the planned discussion with CollyPride on March 24th

  3. Upcoming training sessions Esther will run a plenary session / peer learning session for interviewers before the end of the Quarter (i.e before end March) to gather insights on how interviewing has gone so far, and enable updating of the interviewer learning materials to reflect our context.

Vani will run training in transcription skills as soon as possible - we have budgeted for transcription tasks in this Quarter, so if the training can take place soon, there might be time for people to do some transcriptionwork before the ned of the Quarter - although with token price so low, we will almost certainly have to defer some of the se tasks to next Quarter.

  1. Q2 2025 budget This will be submitted by 10th March. Due to low token price, it is quite minimal, and will focus mainly on transcription tasks. Some of the planned transcriptions we budgeted for in q1 will need to be tranferred to Q2 due to the drop in token price.

We noted that in this Quarter (Q1) we originally budgeted to do 33 new interviews. The drop in token price has meant we have had to reduce this to 15; and in Q2, we will not be able to afford to budget for any new interviews - the WG admin team has agreed it's best to focus on transcriptions this Quarter, so we stay "caught up" and don't accumulate a huge backlog of untranscribed interview material.

Decision Items:

  • Invalid interviews (no consent form, unethical content, etc) is grounds for non-payment. Short interviews (under 30 mins) will be paid on a sliding scale. Expenses will be paid regardless, even if an interview is inavlid or too short.

  • Wrap-up peer learning plenary session for interviewers: looking at good examples, skills exchange etc. Will be led by Esther, probably Wednesday March 19th - afternoon time.

    • [rationale] We need to discuss best practices, what didn't work and other issues

  • Transcription training, led by Vani, will be done as soon as possible - probably 17th March

    • [rationale] so that poeple can undertake the transcribing work that we have budgeted for this Quarter.

Action Items:

  • [action] CollyPride to do an Ambassador Event (think/tank) Discussion "AI Autonomy: Fear, Embrace or Ignore?" on March 24th [assignee] CollyPride [due] 24 March 2025 [status] todo

  • [action] WG Sync Call: Let them know about the Ambassador Discussion with CollyPride on March 24th [assignee] CallyFromAuron [status] todo

  • [action] Kenichi - has he done his survey sharing? Love will follow up with him. [assignee] LadyTempestt [due] 10 March 2025 [status] todo

Keywords/tags:

  • topics covered: interviewing, transcription, Q1 2025 quarterly report, Q2 2025 budget, interview validity, anonymity, training, consent form, expenses, peer learning, token price, knowledge of AI, language barrier

  • emotions: Collaborative, Casual, Friendly

Tuesday 4th March 2025

Governance Workgroup

Narrative:

Advance and Clement have drafted a report on the ETHiopia conference, which they will share in the Open Governance meeting this Thurs (6th March).

Dates for forthcoming sessions:

  • the retrospective on the Q2 decision will be on Tues 1st April

  • initial meeting on budget minimums on Thurs 3rd April. Vani, Peter, Sucre and Effiom to plan it; Vani to make a Discord group to discuss it.

Other possible sessions for future Governance meetings:

    1. We noted that Education Guild are planning to take on some educational sessions on governance issues - Vasu, Lord KIzzy and Vani are planning the sessions now, covering the core governance mechanisms in the Program - so this will cover the longstanding agenda item we have had about whether Governance WG itself should be undertaking governance education sessions

    1. Kevin suggested a meeting with Mindplex - we decided this might not make so much sense as a Governance session, so Kevin will organise this outside of Governance meetings, perhaps under Marketing.

    1. Tevo suggested discussing contribution systems, and who contributes - what are the patterns in the data we have, for example who are the "high cost individuals", and should we aim to decentralise further on fund distribution? Others said that perhaps it's OK if "high-cost individuals" are those who are prepared to do the most work. Also - look at the ratio of amount of work (in terms of MINS) vs amount earned in AGIX

    1. we do need to discuss how to raise the issue of if/when a WG's outputs are felt to be low-quality. What should the pathway be for raising it?

End of Year Report: anything still to be added should be done by this Friday (7th); then Vani will proofread over the weekend.

Discussion Points:

  • Calculating Core Contributors for Q2

  • Signing off Q2 consent form

  • Ideas and dates for future Governance sessions (including Mindplex/outreach to the rest of the ecosystem, Governance education; Contribution systems)

  • End-Of-Year report 2024

Decision Items:

  • We agreed it's time to calculate who is a Core Contributor for Q2 - Tevo will do it later today and post the results in the Governance Discord channel

    • [rationale] allows time for people to nominate anyone who doesn't quite meet the criteria

    • [effect] mayAffectOtherPeople

  • We signed off the Q2 decision consent form

    • [effect] mayAffectOtherPeople

  • We agreed we need to discuss how to raise the issue of if/when a WG's outputs are felt to be low-quality.

    • [rationale] Because it's an issue that needs to be discussed, but it can be sensitive, and there is no clear pathway for how to bring it up.

    • [effect] mayAffectOtherPeople

Action Items:

  • [action] Tevo to calculate who is a Core Contributor for the coming budget decision [assignee] Tevo [due] 4 March 2025 [status] todo

  • [action] Vani, Peter, Sucre and Effiom to plan initial session on budget minimums on 3rd April; Vani to make a Discord group to discuss it. [assignee] Effiom, PeterE, Sucre n Spice, CallyFromAuron [due] 3 April 2025 [status] todo

  • [action] End of Year report - All to add any outstanding material by this Friday (7th) [assignee] all [due] 7 March 2025 [status] todo

  • [action] Vani will proofread end-of-year report over the weekend [assignee] CallyFromAuron [due] 10 March 2025 [status] todo

  • [action] Kevin to organise meeting with Mindplex outside of Governance meetings, perhaps under Marketing. [assignee] TheFreysDeFi [due] 14 March 2025 [status] todo

Keywords/tags:

  • topics covered: Core Contributors , Retrospective planning , budget minimums, ETHiopia event, consent form, consent process, mindplex, Collaboration, contribution, high-cost individuals

  • emotions: low attendance, decisive, forward-looking, Discursive

Wednesday 5th March 2025

Archives Workgroup

Decision Items:

  • Last meeting, we said that we would review, in this meeting, whether or not we can afford to implement the proposed initiative for this Quarter to create a skilled documentation team.

Unfortunately token price is even lower than it was, so we definitely cannot do it - but we noted that the need for it might be obviated by our proposed experiment with not correcting summaries in Q2.

We agreed to close the Issue https://github.com/SingularityNET-Archive/SingularityNET-Archive/issues/235

  • [rationale] We had this in our Q1 budget: "If token price rises sufficiently above $0.55" (the exchange rate at which Q1 budgets were calculated). Price has dropped.

  • [effect] mayAffectOtherPeople

  • Our Quarterly Report draft https://github.com/SingularityNET-Archive/SingularityNET-Archive/issues/231 and Q2 budget draft https://docs.google.com/spreadsheets/d/1BLCkYgyV0rnt-Ntlqp_DXGCm1t3KieAVMS_HaUamqQI/edit?usp=sharing were signed off

    • [rationale] Needs to be submitted on 10th March

    • [effect] affectsOnlyThisWorkgroup

  • André, tool development: see https://github.com/SingularityNET-Archive/SingularityNET-Archive/issues/248 This month the Autosave feature has been completed, and bug fixes have been done

    • [effect] mayAffectOtherPeople

  • We noted that whilst during this Quarter, André has done additional development work as well as the Summary Tool redesign, in Q2 his focus will shift purely to the tool redesign

    • [effect] affectsOnlyThisWorkgroup

  • AI workflow, Q1 - see https://github.com/SingularityNET-Archive/SingularityNET-Archive/issues/233

We noted that Stephen was planning to look this month at approaches to auditing graph data - but this turned out to be a much bigger chunk of work than would fit into this budget line, so it has become a proposal in the BGI Nexus funding round https://deepfunding.ai/proposal/ethical-ai-auditing-a-practice-based-approach/

Instead, in this issue, he will be focusing on building infrastructure tools.

  • [effect] affectsOnlyThisWorkgroup

  • LLM development (see https://github.com/SingularityNET-Archive/SingularityNET-Archive/issues/244 (Feb) and https://github.com/SingularityNET-Archive/SingularityNET-Archive/issues/252 (March) Stephen has built a GitHub App Token Generator - basic proof-of-concept, which uses GraphQL to interface with GitHub data via GitHub API. In March, he will be looking at other data tools that might be useful in our work in Q2.

    • [effect] affectsOnlyThisWorkgroup

  • WG Sync call at the end of March - we decided to reiterate that meeting summaries will not be audited during Q2, and suggest a time limit to submit a summary, after which it will be recorded as "no summary given"

Suggestions ranged from a day to a Quarter - we settled on a month after the meeting took place.

  • [rationale] It shouldn't be turned into a competition; also, some WGs use meeting summary-writing as an onboarding task, so there needs to be time for support to be given and corrections to be made.

  • [effect] mayAffectOtherPeople

  • Research on not correcting meeting summaries in Q2:

We noted that our planned approach might influence what kind of data we get in a meeting summary, and in the resultant knowledge graph.

Hypothesis: we will get well-defined edges (i.e. connections between entities), because this often depends on simple structured data (who was present in a meeting, which meetings a given topic was discussed in, etc); but we might have less information about what anyone said when they discussed a topic, since people in the Ambassador program have historically been less good at documenting this kind of information.

Analysis of what kind of data we are getting with uncorrected summaries might later feed in to the design of the new summary tool - perhaps by attempting to design it to help people collect the kind of data that is missing, if any.

Action Items:

  • [action] Vani has approached DeepFunding, to ask if the Events Circle could take on documentation of DF Town Hall. Focus Group agreed this would be valid - waiting for a reply

Also some interest in Focus Group re: how the Archives tool could be used for further documentation within DF [assignee] CallyFromAuron [due] 2 April 2025 [status] in progress

  • [action] André has fixed the issue with tags where people are copy-pasting strings of comma-separated tags into the summary tool, but they are not rendering as separate tags, but as strings. [assignee] André [due] 5 March 2025 [status] done

  • [action] André and Vani to meet to wrap up the tag taxonomy issue by deciding on an initial controlled vocab for topic and emotion tags [assignee] André, CallyFromAuron [due] 2 April 2025 [status] todo

  • [action] Andre to add "All", "Alfred Itodele" and "Maxmilez" as new Names in the names controlled vocab [assignee] André [due] 19 March 2025 [status] todo

Keywords/tags:

  • topics covered: graph database, Neo4j, summary tool, autosave, decision tracking, Robert Stalnaker, AI ethics, Ai recordkeeping ethics, graph data modelling, Python, JSON, Decentralisation, Decentralization, representing our culture in summaries, tag taxonomy, names taxonomy, controlled vocabularies, comma-separated tags, semantic similarity, recordkeeping ethics, archives ethics

  • emotions: interesting, wide-ranging, well-attended, slightly combative

Thursday 6th March 2025

Governance Workgroup

Narrative:

Quarterly Report: Guillermo shared the quarterly report draft and posted them in the Governance Discord for review until submission is due on Monday

Outstanding meeting summaries: Guillermo has some outstanding meeting summaries which he still needs to add to the Summary Tool

Retrospective on ETHiopia Pop Up City Conference And Hackathon:

  • Clement shared a detailed report on his and Advance's attendance at this event, covering involvement in events, hackathons, media, and community engagement, highlighting challenges, support from volunteers, and key contributions from Samuel and Advance.

  • He shared insights on their engagement at the event, highlighting outreach efforts; discussions on AI, blockchain, and decentralization; keynote presentations; KPI achievements, and future plans for sustaining momentum through partnerships and virtual meetups.

  • CJFrankie applauded Clement and Advance for their outreach efforts in Ethiopia and raised concerns about the general lack of awareness of SingularityNET, especially among African audiences. He suggested allocating resources for physical meetups to enhance engagement and proposed that workgroups focus on making SingularityNET tools more accessible. He also requested clarification on whether SNET tools were used during the Hackathon.

  • Advance explained that most hackathons operate in pools, often sponsored by different ecosystems like Mantle, and Ethereum in the context of the hackathon held during the program. Participants build projects using the sponsors' technologies, and those who integrate these tools effectively can receive funding or prizes. If SingularityNET were to sponsor a hackathon, it could incentivize developers to onboard their projects onto the AI Marketplace by offering prize pools for utilizing SNET services.

  • Clement suggested that if SingularityNET had engaged earlier in the hackathon, it could have introduced the AI Marketplace sooner, leading to more participants utilizing its services. He recommended that SingularityNET consider sponsoring such events in the future to encourage developers to build on its platform and integrate its AI tools.

  • Peter noted that SingularityNET is already sponsoring Hack India but emphasized that larger organizations like Deep Funding or the ASI Foundation are better suited for sponsorships. He highlighted that the AI Marketplace is still in beta, with many services as proof-of-concept. With the ASI merger, platforms like Agentverse and ASI Create will integrate, but onboarding developers remains unclear during this transition.

Discussion Points:

  • Q2 Quarterly Report

  • Outstanding meeting summaries

  • Retrospective on ETHiopia Attendance Pop Up City Conference And Hackathon

Action Items:

  • [action] Advance & Clement to write an article capturing experiences from the Ethiopia event and share it with the team. [assignee] advanceameyaw, Clement Umoh [due] 20 March 2025 [status] todo

  • [action] Clement to discuss with the Marketing Workgroup about uploading event content on media channels. [assignee] Clement Umoh [due] 20 March 2025 [status] todo

  • [action] Clement to follow up with collected contacts to explore potential collaborations. [assignee] Clement Umoh [due] 20 March 2025 [status] todo

  • [action] Guillermo to integrate any comments/corrections to Q1 Quarterly report, and submit it and Q2 budget by the deadline of midnight on Mon 10th March [assignee] guillermolucero [due] 10 March 2025 [status] todo

Keywords/tags:

  • topics covered: ETHiopia event, Hackathon, sponsorship, AI marketplace, Q1 2025 quarterly report, Africa, awareness of sNET in Africa, Retrospective, events attendance, Partnership, virtual meetups, sNET tools, making SingularityNET tools more accessible

  • emotions: Accountability, Organised, informative, Discursive, Educative

Last updated