If we want to understand how AI may infiltrate every aspect of society, we need only look to its current role in economic markets. Algorithms can enable price discrimination such that two consumers are charged two different prices on the same website at the same time. AI bots have conducted insider trading and lied about it in research experiments. Now, rental pricing website RealPage has been accused of price fixing. Instead of having a human intermediary facilitate the exchange of company information and the agreement of artificially high prices, an algorithm plays that role. Critics have pointed to RealPage’s binding price decisions for landlords as the hallmark of collusion.
The Department of Justice (DOJ) has filed a lawsuit against RealPage and is joined by California, Colorado, Connecticut, Minnesota, North Carolina, Oregon, Tennessee and Washington. Separately, Arizona and Washington DC’s attorneys general have also filed a suit. Senator Amy Klobuchar has proposed a bill that prevents the use of algorithms to exchange competitively sensitive information and raise prices. Vice President Kamala Harris has recently proposed a ban on algorithm-driven price-setting tools for landlords to set rents.
While these proposals are focused specifically on price fixing, these examples raise a broader concern. The arc of AI is that it is going to increasingly pervade our lives. That’s not a hypothetical - it’s here and happening all around us on a day-to-day basis. If that’s going to happen, it’s critical that safety progress alongside innovation and not fall by the wayside. AI companies must commit to conducting safety evaluations and continue researching critical AI topics such as explainability and alignment. It’s the least they can do to ensure that their products are safe, effective, and reliable.
The Center for AI Policy (CAIP) has a 2024 action plan and full proposed model legislation. We encourage you to visit both for specific policy measures to ensure safer AI.
Analyzing present and future military uses of AI
AISI conducted pre-deployment evaluations of Anthropic's Claude 3.5 Sonnet model
Slower AI progress would still move fast enough to radically disrupt American society, culture, and business