Uncle Sam mulls spying on clouds being used to train AI
- Reference: 1695659409
- News link: https://www.theregister.co.uk/2023/09/25/us_cloud_monitoring_ai/
- Source link:
According to reports, the Biden administration is looking to introduce rules that require public cloud operators to disclose to the authorities when any customer purchases a level of compute resources that come in above a certain as-yet unspecified threshold.
These rules, which the Department of Commerce would enforce, are intended to create a kind of early warning system to let Washington identify potential AI threats ahead of time, [1]Semafor reported, citing unnamed sources said to be "familiar with an upcoming executive order on artificial intelligence."
[2]
As with other areas of advanced technology, it appears the US government is concerned about hostile countries using compute resources to develop advanced AI models, possibly with nefarious purposes in mind.
[3]
[4]
The CEO of AI company Anthropic, which today confirmed [5]investment from Amazon , also warned earlier this year that AI models could give criminals the power to [6]create bioweapons and dangerous viruses , possibly within two or three years.
ChatGPT developer OpenAI [7]advocated earlier this year for an international agency to be in charge of inspecting and auditing AI to ensure the technology is safe and does not represent a threat.
[8]Amazon to sink $4B into AI dev Anthropic, become its cloud provider
[9]A closer look at Harvard and Google's HPC heart research project
[10]Friendly AI chatbots will be designing bioweapons for criminals 'within years'
[11]Washington plans to block Chinese access to AI cloud services
However, it isn't clear if Washington's rules would apply to all cloud providers that offer resources for AI model training, or if the giants of the industry such as Amazon's AWS, Microsoft's Azure, and Google Cloud are considered to be the only clouds with the scale to provide enough resources for the most advanced models.
Semafor also cited concerns that such measures could amount to a surveillance program if not implemented carefully. For example, many research and development programs [12]using the cloud for high performance compute (HPC) resources could easily trigger the reporting requirement and end up being monitored by the US government.
[13]
We asked the Department of Commerce if it can comment on the formation of a cloud threshold reporting scheme, and will update this article if we get an answer.
China, the bogeyman for the US administration, was not specifically mentioned in Semafor's report, instead it offers the example of a theoretical company in the Middle East building a powerful large language model. China already has substantial cloud resources of its own, provided by companies such as Alibaba, Huawei, and Tencent.
That hasn't stopped Washington from considering blocking Chinese access to American cloud providers, as we [14]reported earlier this year. The US was already said to be looking to enact rules that would require domestic cloud operators such as the big three to obtain explicit approval before allowing customers in China access to services used for training AI models.
[15]
The US was also said to be aiming to [16]tighten up the existing export controls in place that are designed to limit China's access to powerful processors and associated technologies that could be used in the development of advanced AI.
Washington announced just last week that it had finalized carefully planned measures or " [17]guardrails " designed to ensure that funds from the roughly $52 billion CHIPS Act do not end up benefiting companies linked with China or other nations of concern. ®
Get our [18]Tech Resources
[1] https://www.semafor.com/article/09/22/2023/white-house-could-force-cloud-companies-to-disclose-ai-customers
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2ZRIDCvMgRGkWJeoNiBCQnwAAAIM&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZRIDCvMgRGkWJeoNiBCQnwAAAIM&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[4] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZRIDCvMgRGkWJeoNiBCQnwAAAIM&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[5] https://www.theregister.com/2023/09/25/amazon_to_sink_4bn_into/
[6] https://www.theregister.com/2023/07/28/ai_senate_bioweapon/
[7] https://www.theregister.com/2023/05/24/openai_superintelligence_global_agency/
[8] https://www.theregister.com/2023/09/25/amazon_to_sink_4bn_into/
[9] https://www.theregister.com/2023/08/18/google_harvard_hpc_heart_research/
[10] https://www.theregister.com/2023/07/28/ai_senate_bioweapon/
[11] https://www.theregister.com/2023/07/05/washington_block_china_ai/
[12] https://www.theregister.com/2023/08/18/google_harvard_hpc_heart_research/
[13] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=4&c=44ZRIDCvMgRGkWJeoNiBCQnwAAAIM&t=ct%3Dns%26unitnum%3D4%26raptor%3Dfalcon%26pos%3Dmid%26test%3D0
[14] https://www.theregister.com/2023/07/05/washington_block_china_ai/
[15] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_software/aiml&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=3&c=33ZRIDCvMgRGkWJeoNiBCQnwAAAIM&t=ct%3Dns%26unitnum%3D3%26raptor%3Deagle%26pos%3Dmid%26test%3D0
[16] https://www.theregister.com/2023/07/28/us_china_ai_chip/
[17] https://www.theregister.com/2023/09/23/commerce_dept_us_chips_funds/
[18] https://whitepapers.theregister.com/
Re: You Have The Right To Privacy, After We Approve The Content.
Ok, but you're a few years too late. Once upon a time the US announced Total Information Awareness and tried to ram the Clipper chip down our throat. That didn't end well because they did it publicly.
So, they resumed the sort of backdoor policy that they started with Crypto AG, but MUCH more blatant: enter Google and Microsoft. Google sits at the front foor of the web to the point that most end users I know enter URLs into the search field instead of the browser's URL line like well trained puppies.
If you pay attention to how much data and communication is exported from a company by either via Google Docs, MS Office et al you'll see that there is already plenty access to corporate information and IP, but then someone dreamt up a term for hosting data outside so vague that executives could pretend they understood it: "The Cloud" and nebulous it was. Most C level people have absolutely no clue what it means (and telling them it means "someone else's computer" wasn't enough of a hint) and so we collectively made frankly industrial grade espionage even easier by effectively giving them the data they want by storing it on their systems instead of our own, and so lose the ability to at least notice when they take some or mount surveillance (in case you want to stop the surveillance at your desk, the software that stops the data export has already been designated malware by Microsoft Defender, so it appears defends MS more than the user).
What? Oh, it's stored on EU based systems? Really? Who do you think helped establish the Cloud Act, and why do you think that is? Why do you think your authentication happens via a Microsoft server and any access to your corporate data always goes via a site that ends on 'microsoft.com' (the 'lite' version of that is Google Docs)? Have you not noticed they're trying to impose the need for users to establish personal accounts when trying to back up their company 2FA credentials (no, really, just try to back up your MS Authenticator COMPANY 2FA credentials without a personal account).
You know these US crime movies where they can find out personal details about anyone on the planet?
They actually built it. It's here, now.
Lie detector
The thing is you can now get any government announcement, policy thrown into the AI grinder and it will neatly summarise all instances of gaslighting, manipulations, inconsistencies, logic errors and so on.
Governments for years successfully avoided teaching how propaganda works, so that they could use all those techniques on unsuspecting voters.
Now the voter don't need to learn that as the AI can easily do that for them.
The prospect of that is scary for governments. The greatest weapon at their disposal is on the brink of being disarmed.
Re: Lie detector
If they're that worried, I'd half expect them to ban AI technology altogether. With the exception of military applications, of course. Much like they tried to do to cryptography back in the 80s and 90s, really. I mean it worked so well then... (where's the eye roll icon...)
"If they're that worried, I'd half expect them to ban AI technology altogether."
I'm sure they would love to. But that wouldn't be possible - a ban doesn't mean the stuff doesn't exist, just that it's illegal to use / posess etc. Criminals don't care about that, and there will be plenty of places to get it outside of western jurisdiction.
Meanwhile, said jurisdictions will be making and using AI, gaining experience and commercial / military advantage, villains will be obtaining and using it for nefarious purposes, and the west won't have much, if anything to counter it with.
If meddle they must - and they're politicians so they will - their only option is to try to control and regulate. It won't make much difference, of course. It won't be too long before enthusiasts are training models in their garden shed - good luck regulating that!
You Have The Right To Privacy, After We Approve The Content.
Effing NSA, GCHQ and Five Eyes in general. They want access to everything because it may be a "threat" one day.
Say anything, do anything, just get access. Skim the best ideas that can make us profit, delete the things we don't like and persecute radical threats like journalists, privacy advocates and members of the EFF.