AI poses some difficult policy questions
Artificial intelligence news is everywhere. From the recent cover of Time proclaiming, “The AI Arms Race is Changing Everything,” to an AI-generated art on the cover of Vogue, the AI fever is palpable.
While today's focus is on the mind-boggling features of new AI tools, future headlines will center on the inevitable battle to regulate and control the technology. When that time comes, you can bet California's politicians will want to lead.
The hype is warranted. New generative AI tools represent a serious leap in capabilities by generating shockingly coherent text and images. The most wellknown is ChatGPT, a chat interface that has become a cultural phenomenon and the fastest app in history to reach 100 million users.
However, ChatGPT is just the tip of the iceberg. Hundreds of other generative AI tools are being created to create digital art, and being trained to offer more specialized, industry-specific services.
Silicon Valley has gone into a frenzied overdrive as all the major tech giants work to surpass each other's AI capabilities and new startups work to capitalize on the buzz. Wall Street executives have already declared generative AI the technology of the decade, and Microsoft's CEO promised that “AI will fundamentally change every software category.”
While AI has the potential to be the most exciting technology of the decade, it would almost certainly also be the most politically controversial.
Politicians will care about AI because it promises to pour gasoline onto existing tech battles while also introducing a whole new set of problems. For example, the misinformation campaigns around the 2016 election will look like child's play once bad actors implement AI tools and can generate an endless supply of hateful, human-like content.
AI technology will also make current debates around privacy and data rights look quaint. Generative AI allows users to create compelling variations of past art, speech and writing, blurring the lines of plagiarism and intellectual property. AI makes it easy to mimic someone's tone or appearance to create imitation text, speech and even deepfake porn using people's public information.
Regulators will certainly want to see new rules about which data AI models are allowed to be trained on and what they're able to create. But who will draw the line on what is defined as acceptable use and how a person's data can be used? More concerningly, how will any rules possibly be enforced when models are cheap, widely deployed and beyond the reach of our current enforcement tools?
In addition to making existing problems worse, policymakers will also be concerned about the impact on jobs. A recent survey of 1,000 business leaders found that roughly half of those who used ChatGPT have already replaced workers. Generative AI tools are already capable of performing entrylevel tasks. While this doesn't mean we'll see mass white-collar unemployment, these tools could result in an economic restructuring where there's less demand for lower-end audio engineers, copywriters, software engineers, business analysts and any other job done at a keyboard.
How will states like California react if a portion of knowledge worker jobs disappear? Those people will likely find new work in the long-term as AI creates new opportunities, but could lead to unprecedented labor market shocks that destabilizes communities and our tax base.
Whenever these battles arrive, state policymakers need to be among the first to act. Just as with other divisive national issues, California will want to assert its ability to shape the national policy agenda just like it has on carbon emissions, electric vehicles and data privacy protection laws.
California has never shied away from being a leader, and there's no reason to believe this time will be different as Washington D.C. remains gridlocked. While we can't predict how society will react to AI, the early battles will likely be fought in Sacramento.