US AI Legal Decisions: Will They Matter?

January 10, 2023

I read an interesting essay called “This Lawsuit against Microsoft Could Change the Future of AI.” It is understandable that the viewpoint is US centric. The technology is the trendy discontinuity called ChatGPT. The issue is harvesting data, lots of it from any source reachable. The litigation concerns Microsoft’s use of open source software to create a service which generates code automatically in response to human or system requests.

The essay uses a compelling analogy. Here’s the passage with the metaphor:

But there’s a dirty little secret at the core of AI — intellectual property theft. To do its work, AI needs to constantly ingest data, lots of it. Think of it as the monster plant Audrey II in Little Shop of Horrors, constantly crying out “Feed me!” Detractors say AI is violating intellectual property laws by hoovering up information without getting the rights to it, and that things will only get worse from here.

One minor point: I would add the word “quickly” after the final word here.

I think there is another issue which may warrant some consideration. Will other countries — for instance, China, North Korea, or Iran — be constrained in their use of open source or proprietary content when training their smart software? One example is the intake of Galmon open source satellite data to assist in guiding anti satellite weapons should the need arise. What happens when compromised telecommunications systems allow streams of real time data to be pumped into ChatGPT-like smart systems? Smart systems with certain types of telemetry can take informed, direct action without too many humans in the process chain.

I suppose I should be interested in Microsoft’s use of ChatGPT. I am more interested in weaponized AI operating outside the span of control of the US legal decisions. Control of information and the concomitant lack of control of information is more than adding zest to a Word document.

As a dinobaby, I am often wrong. Maybe what the US does will act like a governor on an 19th century steam engine? As I recall, some of the governors failed with some interesting consequences. Worry about Google, Microsoft, or some other US company’s application of constrained information could be worrying about a lesser issue.

Stephen E Arnold, January 10. 2023

Comments

Comments are closed.

  • Archives

  • Recent Posts

  • Meta