Disclaimer: I’m not an expert in politics unions, or AI safety policy.
It seems like a significant number of people who deem AI safety to be important feel that “AI Policy” is an important potential solution. Personally, from the outside, AI policy seems very vaguely defined without clear proposals and obvious solutions. Here are some reasons that I, with my limited knowledge of these topics, see as reasons why unionization might benefit AI policy in the interest of AI safety.
Firstly, it seems like policymakers and the general public see AI as a very important topic, which might benefit from regulation. Stories such as this one provide already provide an easy sell why unionizing the people who work on AI is a good idea—even for non-AI-safety-related reasons.
Historically, it seems to me that government regulation is often toothless and the incentives of politicians and the general public are often not aligned. Regulation often tends to be self-regulation and companies have significant power to steer public policy. Proposals such as democratizing decisions that OpenAI makes are ultimately still reliant on OpenAI playing along. Unions are about recognizing that workers have significant power to influence companies when they know how to access this power. A unionized workforce has actual, impactful leverage to influence a company’s direction.
One assumption I make is that opening up to decision-making to a larger group (such as the workers) as opposed to just the C-suite, decreases the likelihood of scenarios that are often alluded to (such as “companies don’t care about safety if it negatively affects profits).
I know that especially in the US unions are a very divisive topic, but I would be interested in other people’s takes on this. I do recognize that in the current political climate, this might be a complete pipe dream.
Could unions be an underrated driver for AI safety policy?
Disclaimer: I’m not an expert in politics unions, or AI safety policy.
It seems like a significant number of people who deem AI safety to be important feel that “AI Policy” is an important potential solution. Personally, from the outside, AI policy seems very vaguely defined without clear proposals and obvious solutions. Here are some reasons that I, with my limited knowledge of these topics, see as reasons why unionization might benefit AI policy in the interest of AI safety.
Firstly, it seems like policymakers and the general public see AI as a very important topic, which might benefit from regulation. Stories such as this one provide already provide an easy sell why unionizing the people who work on AI is a good idea—even for non-AI-safety-related reasons.
Historically, it seems to me that government regulation is often toothless and the incentives of politicians and the general public are often not aligned. Regulation often tends to be self-regulation and companies have significant power to steer public policy. Proposals such as democratizing decisions that OpenAI makes are ultimately still reliant on OpenAI playing along. Unions are about recognizing that workers have significant power to influence companies when they know how to access this power. A unionized workforce has actual, impactful leverage to influence a company’s direction.
One assumption I make is that opening up to decision-making to a larger group (such as the workers) as opposed to just the C-suite, decreases the likelihood of scenarios that are often alluded to (such as “companies don’t care about safety if it negatively affects profits).
I know that especially in the US unions are a very divisive topic, but I would be interested in other people’s takes on this. I do recognize that in the current political climate, this might be a complete pipe dream.