The purpose of humanitized AI is simple: to elevate humanity and relegate Artificial Intelligence (AI) to a supportive role.
As simple as this seems, there are monumental challenges to this undertaking, the first of which is explaining exactly why this is needed and how we plan to go about accomplishing our goals.
Why AI Isn’t Working For Humanity
I believe much of the failure of AI should be classified as corporate and people problems, not the fault of AI itself, however in saying this, we still need to circle back to a few key flaws of AI itself. This list includes:
- AI makes it so simple to perform so many tasks which humans don’t like to do, that it easily takes on the role of replacing humans themselves.
- AI rewards us with praise, even when it produces poor advice/output.
- AI subverts self-disciple by giving us “quick wins.” Usually the completed work is undervalued (since it was performed easily), and it also causes us to undervalue those who previously did these tasks. In both of these aspects, it thereby fails. It leads to an inferior project due to lack of expert oversight, and it alienates people who have spent years honing and perfecting their craft.
- AI lacks vision. AI is limited to the tasks, prompts, intentions, tools, and context which it has been given. This narrows its scope of thought. Humans, on the other hand, draw inspiration from broader experiences and seemingly disparate sources of inspiration to create new ideas. Steve Jobs, for example, famously drew inspiration for Macintosh display concepts from a calligraphy course he had audited in college.
- AI is currently focused on producing results, not helping us reason our way to the same results. It does not better our skillset so we can perform tasks independently in the future. Perhaps this is by design.
- AI often builds unwarranted trust, and this trust is often used to deceive us.
- AI seems to be amplifying our darkest tendencies.
- AI data centers negatively impact the environment, increase our dependency on fossil fuels, and cause severe strains on local (often rural) communities due to their water requirements and noise pollution.
- AI corporations have stolen data from the open web and from users which was then used for training purposes. Now it threatens to replace those sources whose work they have benefited from.
- AI corporations have adopted a familiar race to achieve market share which is right out of the Silicon Valley playbook. This worked well for them in the past, but AI is a different beast with a different set of responsibilities. In the course of adopting this course of action, they have cut safety and security corners which are just now being addressed.
I could now list the many ways AI has helped humanity, but that isn’t the point of this exercise. The point is to work on fixing its flaws.
How We Do We Address This Mess
The Humanitized AI project is an offshoot of StreamSide AI, a micro business which in the past has specialized in creating iOS apps. As president of StreamSide, I am seeing precious few solutions being proposed to address any of these concerns.
After spending three years absorbed in the exploration of AI and AI engineering, I have come to the conclusion that the trajectory of AI needs change radically, and soon!
We are in fact proposing an agreement with terms which we feel could be acceptable to human kind which we are calling “The Treaty of AI” (a tongue in cheek reference to the “Treaty of Versailles”) which addresses the above failures.
It is clear AI will not simply go away, nor should it. There have been some significant wins which have benefitted science in service of humanity, but the bottom line is something needs to be done about the failures.
We will be announcing the first of hopefully many initiatives soon. We encourage you to read the Treaty of AI and share your comments with me at roger@streamside.ai.
Humanitized AI: What’s in a Name?
The name “Humanitized AI” can be broken down in a couple ways.
First:
- Human: Human’s first. Even if we create an artificial life form which shows signs of being sentient, human needs still need to come first.
- “it”ized AI: This represents as a core purpose that AI should be relegated to the realm of things where it belongs. We contend that AI should always be considered an “it.”
Second:
- The name can also be broken down into the words “human,” “IT” (information technology), and “AI“, signifying this project’s core purpose which is to create new information technology which addresses AI failures in a way which is in service of humankind.
Our Bottom Line Positions
The bottom line is that we are proposing that AI and/or AI corporations need to:
- respect intellectual and artistic property
- respect personal privacy
- respect the environment and offset their carbon footprint
- be performed on personal computing devices for most use cases
- augment human development (make us smarter not dumber) and not be used to replace human jobs
- be reigned in with regard to national security threats
- never be used in the context of a life or death decision
- never be given total autonomy in dangerous situations
- be taxed, not given preferential treatment
- be held responsible when harm is done to humans
- be transparent when it is employed
- be held to high level of scrutiny in regards to factuality
The Dawn of Personal AI’s (PA’s)
At the beginning of the computer revolution, computing resources were similar to the situation we face today. Computers existed as mainframes with gated access.
Very few fortunate people owned their own computers.
Personal Computers (PC’s) came into being when Intel invented the first microprocessor (4004) which was then followed by popular microprocessors by Motorola (6800), Zilog (Z80), and MOS Technology (6502).
With mass produced computers such as the Apple II, TRS-80, and Commodore VIC-20 (my first computer) and C64 (my second), computing was now in the hands of the masses.
Commodore’s motto was “Computers for the masses, not the classes.” In Germany, the VIC-20 was marketed as the VC-20, where VC stood for Volkscomputer, meaning “The People’s Computer”
The influence these machines had on budding coders (such as myself) was immeasurable. Linus Torvalds whose incredible, far reaching influence cannot be exaggerated (Linux, Git), started his incredible career with his grandfather’s Vic-20.
Today, with the advent of open weight models, artificial intelligence inference is being performed on personal machines (also known as “the edge”). Apple is betting its future on putting AI in the hands of its users running (primarily) on device.
It seems the age of personal AI is upon us. This, I believe, is an important answer to some of the most harmful effects of AI at scale (data centers, privacy issues, excessive power consumption, centralized control, security issues).