AI regulation in the US

With so much going on in AI, there has been talk of how it will be regulated. I’m starting this thread about some proposals for regulation in the US.

In California where much of the tech is originated, there are numerous proposals for AI regulation. Most of it has an uphill climb with the powerful corporations in the State.

I’m curious what legislation the corporations themselves want to see. I’ve read more than once that the head of big corporations want to see some guard rails around AI. What do they want these to be?

Here’s some regulations from the Federal government put out by Biden.

Fine words butter no parsnips … :person_shrugging:

The “internet” is still not comprehensively regulated.

Who Regulates the Internet?

The one in charge of regulating the Internet is the Government. Each country will have its own Internet laws, which is why it is easier to gain access to content and information in some countries compared to others.

In the United States, for example, there are several agencies that regulate the Internet. Some of these are:

  • Federal Trade Commission (FTC)

  • Federal Communications Commission (FCC)

  • National Highway Traffic Security Administration (NHTSA)

These aren’t the only agencies of the Government responsible for that. Each agency regulates over their own intellectual properties based on the country’s rules.

This is why the Internet in Europe looks so much different from the Internet in Korea, for example. All governments have their own laws.

As for AI, the genie’s out of the bottle and ahead of the curve … so there’s no chance of regulation, except in countries which impose state censorship.

(1) The full list is 7 pages long … and those are just the outlines.

(2) I’ll bet that those developers will be extremely reluctant to put their “critical information” in the public domain … :face_with_raised_eyebrow:

Presidential executive orders, once issued, remain in force until they are canceled, revoked, adjudicated unlawful, or expire on their terms. At any time, the president may revoke, modify or make exceptions from any executive order, whether the order was made by the current president or a predecessor. Typically, a new president reviews in-force executive orders in the first few weeks in office.

The United States Constitution does not have a provision that explicitly permits the use of executive orders. Article II, Section 1, Clause 1 of the Constitution simply states: “The executive Power shall be vested in a President of the United States of America.” Sections 2 and 3 describe the various powers and duties of the president, including “He shall take care that the Laws be faithfully executed”.

The U.S. Supreme Court has held that all executive orders from the president of the United States must be supported by the Constitution, whether from a clause granting specific power, or by Congress delegating such to the executive branch. Specifically, such orders must be rooted in Article II of the US Constitution or enacted by the Congress in statutes. Attempts to block such orders have been successful at times, when such orders either exceeded the authority of the president or could be better handled through legislation.

Presumably, the CIA will be collecting data on AI developments in other (less regulated) countries and will have to employ the services of the developers of the most powerful US AI systems in order to establish potential threats. Any models thereby created will be in contravention of the “regulations”.

Of course, the most important question is “Who’s going to pay for the implementation of these regulations?”