AFPPresident Biden during an earlier meeting on AI at the White House
NOS Nieuws•vandaag, 18:04
Nando Kasteleijn
editor Tech
Nando Kasteleijn
editor Tech
Makers of the “most powerful AI systems” in the US are required to share safety tests and other “critical information” with the government. This is evident from a presidential decree issued today by US President Biden.
Biden’s decree is part of a series of measures with which the US government is trying to gain more control over AI (artificial intelligence). This specifically concerns companies developing an AI system that poses a “serious risk” to US security. It is not described what kind of risk this involves. In addition, it is not stated who determines when such a risk exists.
A presidential decree is not a law. The president needs Congress for that. Biden therefore calls on parliamentarians to come up with legislation.
Next generation AI
An anonymous government official emphasizes to the Financial Times that the decree will mainly apply to the next generation of AI systems and not, for example, to GPT4, the engine behind the latest version of the advanced text generator ChatGPT.
Although developments are moving quickly, the question is when companies will have to deal with these rules in practice and it suggests that the US does not want to be too harsh on its own companies with these new rules.
The White House and tech giants have been discussing AI for some time. In July, the most important AI players committed to non-binding agreements. This also included the promise that they would share information about the security and reliability of their systems. With Biden’s decision, that lack of obligation disappears.
Watermark for AI-generated work
In addition to requirements for the large tech companies, it is also about developing a watermark so that it is clear when work has been AI generated. There is a limitation here: this will only be used by federal governments, but companies do not have to adopt this. Although the White House hopes to be an example for companies and governments around the world.
It also includes measures to protect Americans’ privacy, promote equality, protect the rights of consumers, patients and students, as well as employees, among other things. The order also ensures that the US government itself uses AI in a responsible and effective manner.
The Biden administration’s moves come on the same day as a voluntary code of conduct for AI initiated by the G7 countries and the EU, Reuters reported yesterday. This sets out how major countries want to deal with AI, in the context of concerns about privacy and security.
AI Summit United Kingdom
A two-day AI meeting will start in the United Kingdom on Wednesday, where its security will be central. Vice President Harris is there on behalf of the US. Committee chairman Von der Leyen is expected to represent the EU and State Secretary Van Huffelen will represent the Netherlands. A delegation from China is also expected.
2023-10-30 17:04:55
#President #Biden #introducing #measures #effectiveness #limited
