About

The Problem

Companies and governments worldwide spent $37.5 billion on artificial intelligence systems in 2019.^[1] That's 4x the entire FBI budget!^[2]

And yet problems abound, even amongst these companies spending so much on cutting edge technology.

Taking down child porn doesn't make money, so businesses will only do it kicking and screaming.
- YouTube allowed child predators to leave comments on children's videos directing others to revealing portions^[3].
- And the AI-created recommendations continued to lead predators to similar videos, months after the initial revelations^[4].
- The same tech companies that spend many millions optimizing ad spend and attempting to create self-driving cars seem to hardly care about preventing child porn. New York Times journalists (not engineers) found illicit images with basic search queries on Bing that flagged Microsoft's own illegal image detection engine^[5].
There's little political gain in pushing through the funding and resources needed to address the problem of online child exploitation from the law enforcement level, so the government can't address the issue well either
- Only about half of the total funding authorized by a 2008 law to fight child pornography is regularly approved, and the total amount is only $60 million^[6]. And markets are getting more difficult for what law enforcement there is to infiltrate.
- The New York Times produced a two-part podcast that covers the scale of the problem and inadequacy of current methods of addressing it^[7]. It's useful listening, but beware - it's also very troubling.

In addition to these issues, consider spending on artificial intelligence as a whole. Who are the top spenders? According to IDC, it's the retail and banking industries^[8]. And you can be certain that their objective isn't to give more money back to you and me. From the IDC report:

"Nearly half of the retail spending will go toward automated customer service agents and expert shopping advisors & product recommendation systems."
"The banking industry will focus its investments on automated threat intelligence and prevention systems and fraud analysis and investigation."

Fraud analysis and investigation is great, but it only stops bad things from happening (doesn't create value, just prevents its loss). And as for retail? More phone menus with automated service is exactly what I want. You too?

Outsourcing and automation has hit physical and manufacturing jobs in developed countries; white-collar jobs may be at risk from AI advances now too^[9].

We can do better.

Deep learning enables never-before-seen scales and efficiencies in both manual labor and knowledge work. Well-built systems can run machines, quality check items, and even create photorealistic images^[10].

Instead of just driving ad money and eliminating human jobs, this power can be wielded for good.

In some places, job loss wouldn't be bad; unlike in elections, fake news and rumors should be used for disruption.

Deep learning enabled content generation could be used to undermine prices and producer revenue, and to direct consumers to help.
AI-created fake news and rumors can undermine trust in illicit markets.

In other venues, artificial intelligence systems can be leveraged to support individuals. We don't know now how this would work. But we do know that it's an important research area that isn't being looked at by any company with a profit motive, nor by people too engrossed by superintelligence-ethics-x-risk sexy thinking.

So let's do these, together.

We're deploying technology to destroy the bad that isn't addressed well by governments or businesses, and publishing strategic research on using AI to build better lives for individuals.

And we need your help.

Making this happen doesn't need to cost anywhere near the $37 billion spent on deep-learning-for-profit. But it does take money. Here's what we need to get off the ground.

Funding Goalposts (2 Year Runway)

Computers and Platforms: $47,500 in computational infrastructure buys machines to run sensitive compute locally.

Salaries and Healthcare: $575,000 gets us off the ground for two years. $2,500,000 would fund an endowment that would let us run indefinitely and balance immediate impact with long-term scale.

Security and Other Operating Expenses: $39,000 pays for security measures, legal expenses, and psychological support needed for working in sensitive areas.

Who We Are

We bring together a unique set of experience in national security, intelligence, finance, marketing, and machine learning. During this initial prototyping-and-fundraising stage we're holding back names for personal and family security reasons. Get in touch for a more personal (and private) conversation.

Contact: CreateAWorld [at] protonmail [dot] com

← Home