Develop, but don’t deploy. No, it is not about keeping up with China.
I encourage you to listen to this podcast from “Your divided attention“. While acknowledging the transformative power of AI, the presentation paints a very dark picture and urges action to limit deployment before it is too late.
The logic of the pessimism is based in the speed-oriented motive of the tech giants to gain market share and the lack of understanding of AI developers of how their tools work or where the use of these tools make eventually take us. The program makes a comparison to social media, another implementation with great promise and still far less powerful than ai, and suggests that the competitive zeal of the tech companies either failed to consider or didn’t care about the negative consequences of their social media services. We should have learned a lesson.
The argument continues to note that the magnitude of the resources that are required to make AI work prevents university researchers who are less interested in financial success and competition from conducting research on what may go wrong and how users may be harmed. There are just no incentives or requirements that the tech industry think and act in a similar way.
What about China? While it is true that China is a significant competitor when it comes to AI development, China is resistant to deploy AI applications for citizen use. They fear what their citizens may learn or do with the tool and since citizen use cannot be predicted tech companies are careful to control use.