“OpenAI is saying that it can help govern AI or evolve ‘security and safety controls.’ However, clearly, as a company that stands to profit from AI adoption, there could be a conflict of interest here. If this partnership program is successful, it continues a trend that we’re seeing away from public sector-supported frameworks to govern technology and toward private-sector best practices,” he said. “We should also consider how seriously other countries will take OpenAI’s claim that it will be an ally in providing democratic AI, something it hasn’t even clearly defined. It makes it clear that its primary partner is the US government. What are other countries that have recently entered into trade disputes or even more serious conflicts with the US to make of that association?”
Jackson felt particularly strongly about where the current AI trends may lead if OpenAI delivers on its stated goals.
“Let’s look at it from the perspective of the services that OpenAI is offering to bring to citizens through partnering with governments. There’s a concept called disintermediation, which examines how technology companies are usurping the relationships that democratic governments have with their citizens by providing the key information and services that citizens historically depended on the state for. What OpenAI is proposing would without a doubt represent a power shift from the state to a private company for a pretty considerable range of informational interactions,” he said. “For example, OpenAI suggests it could provide ‘customized ChatGPT to citizens,’ which would localize language and imbue cultural considerations into the service. The implication is that the partner government would then use this platform to deliver some set of services to those citizens. However, instead of [the government] owning the relationship with citizens, OpenAI captures that.”