WEBVTT

STYLE
::cue {
    color: white;
    background-color: #383943A3;
}

00:00:00.919 --> 00:00:07.080
This presentation explains a development path that
began with Faithful Intelligence Theory as a research

00:00:07.080 --> 00:00:14.599
contribution and is now moving into a governed application
environment through the Anointed EDU AI Companion

00:00:14.599 --> 00:00:15.066
Series.

00:00:16.066 --> 00:00:22.146
The important point is that we are not claiming the
theory has suddenly become validated because we are

00:00:22.146 --> 00:00:23.542
using it in development.

00:00:23.932 --> 00:00:30.652
Instead, the companion program gives us a disciplined
environment in which selected Faithful Intelligence

00:00:30.652 --> 00:00:36.556
principles can be translated into engineering requirements,
controls, evidence, and future tests.

00:00:37.716 --> 00:00:44.756
That creates a bridge between scholarship and application
while preserving the difference between theory, implementation,

00:00:44.756 --> 00:00:46.310
and empirical validation.

00:00:48.350 --> 00:00:54.750
The foundational paper defines Faithful Intelligence
as a form of human capability developed through responsible

00:00:54.750 --> 00:01:00.990
engagement with artificial intelligence, while judgment,
verification, ethical reasoning, practical wisdom,

00:01:00.990 --> 00:01:03.562
and accountability remain under human direction.

00:01:04.562 --> 00:01:08.254
The theory’s central mechanism is Guided Reflective
Verification.

00:01:08.564 --> 00:01:15.364
Rather than simply receiving an AI answer, the learner
evaluates evidence, sources, assumptions, alternative

00:01:15.364 --> 00:01:18.435
interpretations, ethical implications, and consequences.

00:01:19.595 --> 00:01:24.555
The paper also emphasizes Human Agency and a bounded
Cognitive Partnership.

00:01:25.715 --> 00:01:32.435
Those ideas become especially important when we begin
asking how an AI application should be designed, because

00:01:32.435 --> 00:01:38.835
they give us a research-grounded vocabulary for what
human responsibility and responsible AI participation

00:01:38.835 --> 00:01:39.476
should look like.

00:01:41.595 --> 00:01:44.689
There is an important boundary we need to preserve.

00:01:45.929 --> 00:01:52.489
The foundational paper explicitly distinguishes Faithful
Intelligence Theory from prompt engineering, theology

00:01:52.489 --> 00:01:55.917
of AI, and a general theory of AI governance.

00:01:56.917 --> 00:02:04.406
Its explanatory object is human learning and formation
in AI-supported educational environments.

00:02:06.726 --> 00:02:11.235
So our present development work should not rewrite
the theory after the fact.

00:02:13.555 --> 00:02:20.915
What we can do is use selected constructs and principles
from the theory as a design lens for a bounded application

00:02:20.915 --> 00:02:21.606
context.

00:02:23.766 --> 00:02:30.246
That is a defensible research-to-application move:
the theory informs requirements and questions, while

00:02:30.246 --> 00:02:36.326
the technical architecture remains a separate implementation
artifact that can generate evidence back into the

00:02:36.326 --> 00:02:37.376
research program.

00:02:39.496 --> 00:02:43.657
This is the connection that emerged from the companion
development work.

00:02:43.657 --> 00:02:49.126
A research principle can become operationally inspectable
when we translate it through a chain.

00:02:50.286 --> 00:02:55.246
For example, Human Agency becomes a requirement that
human authority remain explicit.

00:02:55.396 --> 00:03:00.916
That requirement can influence technical controls
such as function-bounded companions, authority-aware

00:03:00.916 --> 00:03:06.437
retrieval, escalation, and clear distinctions between
canonical records and AI interpretation.

00:03:07.437 --> 00:03:15.197
Then we collect implementation evidence: logs, provenance,
correction records, validation states, and interaction

00:03:15.197 --> 00:03:15.785
evidence.

00:03:15.935 --> 00:03:21.728
Finally, we review whether the system actually behaved
consistently with the intended principle.

00:03:22.198 --> 00:03:29.318
That last step matters because merely naming an ethical
principle does not prove that a technical control

00:03:29.318 --> 00:03:30.040
is effective.

00:03:31.840 --> 00:03:36.719
The companion series gives us a concrete application
environment.

00:03:36.719 --> 00:03:43.119
Instead of one unrestricted assistant, the architecture
uses specialized companions operating against one

00:03:43.119 --> 00:03:44.425
governed knowledge layer.

00:03:44.815 --> 00:03:49.135
Search and Discovery has one function.

00:03:49.135 --> 00:03:50.844
Citation and Evidence has another.

00:03:52.164 --> 00:03:57.812
Methodology, Learning, Educator, and later companions
each have bounded tasks.

00:03:58.282 --> 00:04:06.362
This architecture aligns naturally with Faithful Intelligence
concepts because it resists the idea that AI should

00:04:06.362 --> 00:04:09.444
become an unbounded substitute for human judgment.

00:04:10.684 --> 00:04:18.684
The companion can retrieve, explain, compare, or recommend
within its assigned function, but canonical institutional

00:04:18.684 --> 00:04:24.585
truth remains outside the model and under governed
human and institutional authority.

00:04:26.545 --> 00:04:33.825
The Faithful Intelligence page on the AI site now
expresses this pathway directly: ethical principle,

00:04:33.825 --> 00:04:41.265
engineering requirement, technical control, implementation
evidence, validation test, interaction evidence, and

00:04:41.265 --> 00:04:42.186
governed review.

00:04:43.506 --> 00:04:47.506
That sequence is important because it makes ethics
traceable.

00:04:47.506 --> 00:04:51.186
Human responsibility can shape authority controls.

00:04:51.186 --> 00:04:55.539
Evidence fidelity can shape provenance and epistemic-status
fields.

00:04:55.929 --> 00:05:00.066
Stewardship can shape retrieval boundaries and data
minimization.

00:05:01.226 --> 00:05:06.453
Discernment can shape uncertainty, escalation, correction,
and human review.

00:05:06.923 --> 00:05:14.283
But the development rule remains strict: the presence
of the Faithful Intelligence layer is not evidence

00:05:14.283 --> 00:05:17.963
that the system is ethically effective.

00:05:17.963 --> 00:05:19.223
Effectiveness must be evaluated.

00:05:21.024 --> 00:05:29.103
The companion program also creates a second connection
to the Faithful Intelligence research agenda: development

00:05:29.103 --> 00:05:32.726
itself becomes evidence-generating activity.

00:05:34.046 --> 00:05:40.126
We record what we observed, why it mattered, what
decision followed, and what changed.

00:05:40.126 --> 00:05:45.325
We distinguish formation provenance from implementation
evidence and validation evidence.

00:05:45.795 --> 00:05:50.675
At the same time, the program has an IP governance
gate.

00:05:50.675 --> 00:05:55.993
Potentially novel computing mechanisms are screened
before detailed intentional public disclosure.

00:05:57.233 --> 00:06:03.553
That is why the public Development Feed can document
the logic of formation while remaining summary-only

00:06:03.553 --> 00:06:04.563
where necessary.

00:06:04.873 --> 00:06:12.313
Patent screening is not a patentability conclusion;
it is simply another governance control over the development

00:06:12.313 --> 00:06:12.991
process.

00:06:15.031 --> 00:06:18.630
This produces a public research-to-application loop.

00:06:18.630 --> 00:06:21.466
The foundational paper remains the scholarly source.

00:06:22.706 --> 00:06:27.666
The companion program becomes a governed application
environment.

00:06:27.666 --> 00:06:31.592
The Faithful Intelligence page explains the connection
publicly.

00:06:32.912 --> 00:06:39.552
This narrated presentation can then become a short
video embedded back into that page and distributed

00:06:39.552 --> 00:06:41.961
through LinkedIn and Anointed Holdings.

00:06:42.431 --> 00:06:48.752
Importantly, those external channels are downstream
distribution, not new sources of authority.

00:06:49.752 --> 00:06:55.032
They should point back to the governed publication
and development records.

00:06:55.032 --> 00:07:00.921
Questions, defects, responses, and interaction evidence
can then become inputs for future research and refinement.

00:07:02.881 --> 00:07:08.720
The larger significance is that Faithful Intelligence
is moving from theory toward testable application

00:07:08.880 --> 00:07:11.416
without collapsing the distinction between the two.

00:07:11.806 --> 00:07:16.206
The theory remains a conceptual scholarly contribution.

00:07:16.206 --> 00:07:21.726
The AI Companion Series provides a bounded environment
where selected constructs can inform requirements

00:07:21.726 --> 00:07:22.573
and controls.

00:07:22.963 --> 00:07:30.403
Development evidence can then show what was actually
implemented, and validation can test what those controls

00:07:30.403 --> 00:07:31.117
actually do.

00:07:31.827 --> 00:07:39.427
That gives us a disciplined cycle from research, to
requirements, to controls, to evidence, to validation,

00:07:39.427 --> 00:07:40.856
and then back to refinement.

00:07:41.246 --> 00:07:47.486
The video created from this presentation will become
another governed public explanation of that evolving

00:07:47.486 --> 00:07:49.319
research-to-application pathway.

00:07:49.709 --> 00:08:02.829
Source basis: Whiteside, J. E. (2026), Faithful Intelligence
Theory (FI-FTP-2026-001, v1.0), Anointed EDU Press;

00:08:02.829 --> 00:08:08.858
Anointed EDU AI Companion development architecture
and governed IP-development track.

00:08:09.328 --> 00:08:17.123
Development/application statements are presented as
formation or design decisions unless separately validated.

00:08:18.443 --> 00:08:25.163
This presentation is also part of the Whiteside Research
Series, connecting the author's scholarly research

00:08:25.163 --> 00:08:29.443
program to the governed application, testing, and
public-evidence pathway.

00:08:30.683 --> 00:08:38.043
The Whiteside Research Series designation identifies
the research lineage; Anointed EDU remains the publishing

00:08:38.043 --> 00:08:41.496
and institutional application environment for this
presentation.

00:08:42.496 --> 00:08:46.256
Learn more about my study at ai.anointededu.com