Aiming AI
at What Matters

How we point the most powerful technology
ever built at people and the planet.

JOEY BARBUSH  .  COM 100  .  PERSUASIVE SPEECH  .  APRIL 2026

ARROW KEYS TO ADVANCE / PRESS F FOR FULLSCREEN
SLIDE 1 OF 10

Imagine an AI that...

◔

Predicts a hurricane 10 days out, in under a minute.

✚

Catches a tumor a tired doctor missed.

✦

Finds a battery material that makes clean energy cheap for everyone.

Not science fiction. Already exists.

SLIDE 02 / THE HOOK
2 / 10

The technology already works.

10 days
accurate forecast in under a minute

Google's GraphCast beats traditional weather forecasting. Exactly what we need for stronger storms and accelerating climate disruption.

2.2M
new crystal structures discovered

Google's GNoME unlocks materials for better batteries, cleaner energy storage, and solar at scale.

Stanford HAI, 2024 AI Index Report
SLIDE 03 / THE PROOF
3 / 10
90%

of the most influential AI models in 2024 came from private companies.

Companies whose primary obligation is to investors. Not to solve the world's problems.

Up from roughly 60 percent the year before.

Stanford HAI, 2024 AI Index Report
SLIDE 04 / THE SCALE
4 / 10

Why is this happening?

Not talent.  The best engineers in the world are American.

Incentives.  Companies answer to investors, so they build what makes money fast.

Result.  AI gets pointed at ads, engagement, content. Not at curing diseases or fixing the climate.

Ahmed, Wahed, Thompson · Science (2023)  .  NITRD AI R&D 2024
SLIDE 05 / THE CAUSE
5 / 10

We have done this before.

$257B
Apollo program in 2023 dollars

What public investment built:

  • GPS
  • Satellite communications
  • Modern medical imaging
  • The technology economy itself

"When America decides something matters, we fund it."

Dreier · Space Policy (2022)  .  CHIPS & Science Act, CRS (2023)
SLIDE 06 / THE PRECEDENT
6 / 10

What public AI investment could fund.

A National AI Research Trust. NASA-modeled. Around 50 billion per year.

◯
Climate modeling and clean energy
✚
Healthcare and affordable medicine
◇
Accessibility tools for people with disabilities
≈
Food and water security
▲
AI for under-resourced schools
✦
Mental health support tools
SLIDE 07 / THE SOLUTION
7 / 10

You don't have to wait. Fork it today.

Real open-source projects. Real GitHub repos. Real contribution paths.

github.com/google-deepmind/graphcast
google-deepmind / graphcast
10-day weather forecasting AI
The model from slide 3. Apache 2.0 licensed. Real climate impact.
★ PythonApache-2.0
github.com/Project-MONAI/MONAI
Project-MONAI / MONAI
Medical imaging deep-learning framework
Used by hospitals and researchers worldwide for tumor detection, segmentation.
★ PythonApache-2.0
github.com/OpenMined/PySyft
OpenMined / PySyft
Privacy-preserving AI for sensitive data
Lets researchers train on healthcare data without ever seeing it.
★ PythonApache-2.0
github.com/google-deepmind/alphafold
google-deepmind / alphafold
Protein structure prediction for drug discovery
200M+ structures. The reason new medicines get developed faster.
★ PythonApache-2.0
SLIDE 08 / THE OPEN PATH
8 / 10

What you can do.

You are the next generation that builds this technology.

  • Build a class project that helps disabled classmates access learning materials.
  • Contribute to an open-source healthcare or climate AI project.
  • Choose a major or career that points your skills at problems worth solving.
  • Use AI for the things that count, starting today.
SLIDE 09 / THE CALL
9 / 10

We fund it.
We build it.
We use it to lift people up.

The next generation that does that, is sitting in this room.

Thank you.

SLIDE 10 / THE CLOSE
JOEY BARBUSH · 2026