Members of the California Legislature are amongst the process of forwarding a new AI Bill dubbed “SB-1047” or “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act”. A new bill that will handle the way AI technologies are developed and force new security and safety protocols for AI developers.
SB-1047: Bill
As of August 22, 2024 page 1, paragraph 2 reads as following:
“This bill would enact the Safe and Secure Innovation for Frontier Artificial Intelligence Models Act to, among other things, 1require that a developer, before beginning to initially training train a covered model, as defined, comply with various requirements, including implementing the capability to promptly enact a full shutdown, as defined, and implement a written and separate safety and security protocol, as specified. 2The bill would require a developer to retain an unredacted copy of the safety and security protocol for as long as the covered model is made available for commercial, public, or foreseeably public use plus 5 years, including records and dates of any updates or revisions and would require a developer to grant to the Attorney General access to the unredacted safety and security protocol.
3The bill would prohibit a developer from using a covered model commercially or publicly, or covered model derivative for a purpose not exclusively related to the training or reasonable evaluation of the covered model or 4compliance with state or federal law or 5making a covered model or a covered model derivative available for commercial or public public, or foreseeably public, use, if there is an unreasonable risk that the covered model or covered model derivative can will cause or materially enable a critical harm, as defined. 6The bill would require a developer, beginning January 1, 2028, 2026, to annually retain a third-party auditor to perform an independent audit of compliance with the requirements of the bill, as provided.
those provisions, as prescribed. The bill would require the auditor to produce an audit report, as prescribed, and would require a developer to retain an unredacted copy of the audit report for as long as the covered model is made available for commercial, public, or foreseeably public use plus 5 years. The bill would require a developer to grant to the Attorney General access to the unredacted auditor’s report upon request. The bill would exempt from disclosure under the California Public Records Act the safety and security protocol and the auditor’s report described above.“
What does this mean?
SB-1047 enforces the following:
1: Requires the developer of an AI technology to introduce a method of shutting it down prior to its proper training and development.
1.B: Include safety measures for the AI technology that ensures it shuts down when the time maybe needed “as specified“.
2: Requires the developer to retain a personal copy of the safety and security protocols that would be required to disable the AI technology whenever the technology is used for commercial or public use of any kind.
2.B: The developer must keep a copy of a record pertaining to the creation, development, revisions, and/or updates regarding to their public/commercial use AI. This record must pertain the past 5 years minimum of the regarded information stated above.
2.C: Requires the developer of a public/commercial use AI to submit a copy of its safety and security protocols to the Attorney General, also including giving the Attorney General access to these safety and security protocols directly.
3: Prohibits a developer from making a covered model (public/commercial use AI or AI in similar fashion) that is not designed primarily for training or “reasonable evaluation” related reasons.
4: Prohibits a developer from making a commercial/public use AI that does not meet the requirements of state or federal regulations.
5: Prohibits a developer from making a commercial/public use AI that is implemented with features that are/can be deemed as “unreasonable risk” towards the public.
6: Beginning on January 1st, of 2028, 2026. Developers of commercial/public use AI will be required to keep a third-party auditor to perform independent audits of the developers legal cooperation and compliance of the new California “Safe and Secure Innovation for Frontier Artificial Intelligence Models Act”.
This is only the beginning of the Bill, for those that wish to continue reading the entirety of the Bill we recommend visiting the official California Legislative Website here: https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240SB1047
Who does this affect?
This bill primarily seems to be focused towards larger tech companies that are utilizing AI in possibly malicious or uncontrollable manners that are based and operating around the State of California. Many big tech companies are based in or operate around the State of California such as the following:
Google, or Googleplex located in 1600 Amphitheatre Parkway, Mountain View, California is the Headquarters for Google Inc. With their headquarters being located in the State of California, Google will be required to implement the following changes for all of their currently implemented AI software, such as but not limited to; Google Assistant, Google Search AI, Google Translate, Bidirectional Encoder Representations from Transformers (B.E.R.T.), and TensorFlow.
Amazon
While Amazon is located primarily in 2121 7th Ave, Seattle, Washington, they still continue business in the State of California, therefore in order for them to continue operating AI based software in the State of California changes would need to be implemented into their AI software programs such as Alexa AI, Amazon Lex, and Amazon Forecast for start. Similiar to Google, Amazon would need to ensure they meet the legal standards stated in the bill above.
Microsoft
Similar to Amazon, while Microsoft is located in Redmond, Washington, this bill would still make business operations within the State of California unlawful unless Microsoft also reform their AI software. This includes but is not limited to; Cortona AI, Microsoft Translator, Azure AI, and Dynamics 365 AI.
What could this mean for Government Involvement?
In terms of Government Involvement and access developers now need to give access to the security and safety protocols of their designed AI’s (if it is to be for public/commercial use) to the State of California Attorney General. As of now, it is unclear if others are given access to this information or if it is the Attorney General of the state itself that has sole access. However, it is also unclear as to what other information could be deemed as a “safety” or “security protocol”.
It’s possible that this could lead to more government oversight and involvement in the development of AI technologies even more so within the future, or it’s possible loopholes could be created from this new bill that allows an Attorney General to state something is a “safety and security protocol” when others may deem otherwise, however this is in no way confirmed at this time.
Where is this Bill now?
As of August 22, 2024, according to the official California Legislative Information website. Bill SB-1047 was read a third time on 05/21/24 and passed with 32 Ayes and 1 Noes and was ordered to the Assembly. As of the release of this article, on August 20th, 2024, it was read the second time in the Assembly Committee and is awaiting it’s 3rd reading before being sent towards the California State Governor for final approval which will take effect on January 1st, of 2028, 2026.
You can view the entire history of the bill SB-1047 Here.
Sources
Legal Sources (Creditable Sources):
State of California’s Legislative Information Website: https://leginfo.legislature.ca.gov/faces/home.xhtml
Official SB-1047 Bill: https://leginfo.legislature.ca.gov/faces/billTextClient.xhtml?bill_id=202320240SB1047
Other Media Sources (Questionable Sources):
Reuters: https://www.reuters.com/technology/artificial-intelligence/big-tech-wants-ai-be-regulated-why-do-they-oppose-california-ai-bill-2024-08-21/
Columbia Journalism Review: https://www.cjr.org/the_media_today/california_ai_bill_journalism.php
The Journal: https://thejournal.com/Articles/2024/08/19/California-AI-Regulation-Bill-Moves-to-Assembly-Vote-with-Key-Amendments.aspx
Un-Recommended Sources (Editable Sources were biasy is possible or likely):
Wikipedia: https://www.wikipedia.org/
Good Morning America (YT): https://www.youtube.com/watch?v=XcO-O5VKg7E
Yahoo Finance (YT): https://www.youtube.com/watch?v=QhO5DhWIp4w
Problems with the article? Suggestions? Or something to add? Contact Seeker on Discord or the Forums to submit a request! Or leave a comment below!