This is the subject of my latest column for Bloomberg, and here is part one:
The security movement almost reached its peak in March 2023 with a petition for a six-month moratorium on the development of AI, which was signed by many luminaries, including experts in the field of AI. As I argued at the time, it was a bad idea, and it got nowhere.
Fast forward to now. Senate Majority Leader Chuck Schumer and his working group on AI have released a directive for federal policy. The plans include more federal support for AI research and development, and a consistent recognition of the national security importance of the US maintaining its lead in AI. Lawmakers seem to understand that they would rather deal with the dangers of US-based AI programs than fight China’s advances without a US counterweight. The early history of Covid, in which the Chinese government behaved recklessly and transparently, has driven this realization.
Just as important is the behavior of the big tech companies themselves. OpenAI, Anthropic, Google and Meta all released major service upgrades this spring. Their innovative services are smart, fast, flexible and efficient. Competition has intensified, and that will spur innovation.
Note this:
The biggest current obstacles to AI development are the hundreds of pending AI regulatory bills in many US states. Many of those bills, intentionally or not, will severely limit the development of AI, such as one example in California that would require pre-approval of advanced systems. In the past, this haste at the state level has often led to the consolidation of the state, so that overall control is consolidated and less complicated. Whether the final positive results from that process are uncertain, but Schumer’s project suggests that the federal government is more interested in accelerating AI than blocking it.
Security work will continue under many considerations, both private and government, but “AI security,” as an intellectual entity, has reached its peak.
This piece was written before some recent disputes at OpenAI, and the debate does not depend on any particular interpretation of those events, one way or the other.
Arnaud Schenk has some helpful explanations, especially for those who can’t read the full column.
Here is a version of the column, adapted from Don Mclean’s song “American Pie.”
Source link