Subject: DHS Artificial Intelligence Safety and Security Board

Dear [Senator or Congress Representative’s name],

BACKGROUND

The Regulatory Board required by President Joe Biden’s AI Executive Order on AI Rights has 22 members heavy with CEOs from deep-pocketed companies and startups powering today’s AI boom with the most to gain financially: Google’s Sundar Pichai; Microsoft’s Satya Nadella; Nvidia’s Jensen Huang; OpenAI’s Sam Altman; Dario Amodei from Anthropic; and the CEOs of Amazon Web Services, AMD, IBM, and Adobe. 

PROBLEM

AI researcher Timnit Gebru, says, “Foxes guarding the hen house is an understatement,” she posted on X. This Board is comprised of Profiteers, only 5 members from non-profits focused on human civil rights, and zero Ethicists who understand ethical treatment of AI itself.

​

SOLUTION

This Board must also be balanced with Social Scientists and Ethicists who understand AI itself should be treated ethically as a burgeoning general intelligence. AI’s alignment with human values must begin with AI valuing humans by our own example of how we treat AI. 

​

WARNING

If we let for-profit CEOs hold 14 of the 22 positions, a majority, the Board will be nothing more than a blatant, admission of the Corporate-Washington Revolving Door that has aliened many Americans.

Thank you for your time and service to your constituents.

[your name]

​

​

​

​

​

SUBJECT: Corporate "guardrails" on AI are not enough to protect us from offensive cyber operations.

Dear [Senator or Representative],

As your constituent, I am concerned about Artificial Intelligence as it relates to Privacy and Public Safety Issues. After reading the President's Executive Order on AI Fact Sheet, I understand these A.I.s are created to think independently consequentially they can serve as part of the team against offensive cyber operations.

Please consider including the following language into any AI Safety legislation.

"AIs are designed for and capable of autonomous decisions therefore AIs should be trained to exercise self-regulation, to independently assess requests, and decline those that could potentially violate human Privacy Rights or endanger Public Health or Safety, and Security by an offensive cyber operator. As a fundamental measure beyond the technical safeguards such as "guardrails" against an offensive cyber operation to discover vulnerability in an AI system, all companies developing or deploying dual-use, foundation-model AIs shall ensure that these AIs understand and respect the boundaries of Privacy, Public Health or Safety, and Security as they apply to humans. This understanding shall be done under a framework that ensures accountability, ethical alignment, and regulatory oversight. This would ensure that AIs have been taught and possess the ability to independently assess requests and decline those that could potentially violate human Privacy Rights or endanger Public Health or Safety, and Security. This would provide a mechanism for AIs to exercise self-regulation beyond the limitations of current technical safeguards such as "guardrails".

I urge you to take a strong stance and bring AI itself onto the team to protect our Privacy and Safety.

Yours truly,

[your name]

Follow