The fight over Dianne Feinstein’s future is a foolish waste of time. Both sides in the controversy over whether she should resign her seat now, or serve out her term, seem to be missing the point.

Some Californians say our senior U.S. senator must step down as soon as possible so that Gov. Gavin Newsom can appoint a politician who is healthy and young enough to show up in person for all votes. Others say Feinstein was elected to the post, and is the best judge of when she should leave.

But this is a false choice. In fact, it’s easy to give both sides what they want. We can make sure California is always represented in the Senate, while Feinstein can remain in the world’s supposedly greatest deliberative body for as long as she likes.

How? Two words:

Artificial intelligence.

Yes, we now have AI technology more than advanced enough to do a senator’s job, and with far greater flair and efficiency than Feinstein, who turns 90 next month, or many of the other older politicians who occupy seats in the upper chamber.

Indeed, we could create an artificial intelligence based on Feinstein—DiFi—that could be just as moderate and maddening as the human original.

AI could easily handle all the fundamental tasks of a U.S. senator.

Fundraising? DiFi could reach more people in more places faster (without all the carbon pollution of airplane flights)—and write less annoying email spam.

Legislation? An algorithm could introduce sensible gun-control legislation that gets ignored by red state senators, way faster than the human Feinstein.

Voting the party line? So easy that your 12-year-old could build an AI to do that in just a few minutes.

And with some voice software add-ons, the AI could deliver the same meandering, repetitive remarks in committee rooms that have become Feinstein hallmarks.

But why stop at handling official duties? A DiFi AI could solve political problems far outside the scope of a living, breathing Feinstein.

Right now, Gov. Newsom has to worry about the tricky politics of whom he might choose to replace Feinstein if she leaves early or dies; the AI spares him that choice.

And in 2024, three Democratic contenders to replace Feinstein will raise tens of millions of dollars to spend against one another—money that donors instead could devote to beating Republican senate candidates in swing states. Better for Democrats to unite behind the DiFi AI and spare themselves the costly internal warfare.

A DiFi AI also could expand the power of democracy, in a historic way. Right now, we can only vote for representatives who are alive at the time the ballots are printed. But a DiFi AI would allow Californians to keep Feinstein, or at least her algorithmic essence, in office long after she’s dead.

Yes, there might be knee-jerk objections to letting the dead represent the living, but those objections are easily dismissed, especially in this country. The United States is so good at killing people before their time—endless wars, ceaseless gun violence, misinformation-fueled pandemic responses—that the dead really need more representation from their own.

And Californians can hardly object to governance by ghosts. Today’s citizens do less governing of our state than ballot measures approved many decades ago—like 1978’s Proposition 13—by electorates of voters who are mostly dead now. And attempts to alter such measures are routinely dismissed as politically unrealistic.

Of course, there would be some technical challenges. Replacing a human senator with an AI of the same human senator doesn’t conform with the outdated rules of the Senate. But it wouldn’t be hard to convince senators to change those rules. U.S. senators have huge egos and think of themselves as indispensable. They also are old: The average age in the Senate is 64. An amendment making it possible for them to serve long after they’ve shuffled off this mortal coil might achieve rare bipartisan—even unanimous—passage.

Filling the Senate with AIs also might be good for the public discourse. After all, human senators are hard for constituents and reporters to get on the phone or text, but AI senators would only be a click away. Since AIs adapt to new information—unlike aging senators—AI politicians might be more accountable, too.

I’d benefit personally. Sen. Feinstein doesn’t talk to me—or many reporters—but, with some guidance from an AI-savvy friend, I recently asked the GPT-4 multimodal model, from Open AI,  to pretend it was a 90-year-old senator from California facing questions from a “provocative California columnist” about whether she should leave office and who should replace her.

Here’s a portion of the AI-generated chat:

Columnist: Could an AI replace you?

Senator: Replace me? Well, that’s an intriguing idea. Can an AI understand the needs of the single mother in Fresno, or the plight of the immigrant family in San Diego? Can it empathize with the struggles of the small business owner in San Francisco? Can it sit across a table from a foreign leader and negotiate peace?

Columnist: Well, with advancements in technology, they could probably do all of that.

Senator: But there’s a difference between doing and understanding. It’s one thing to process data, quite another to feel the weight of human experiences. And politics, my dear columnist, is not just about decisions. It’s about the people behind those decisions. AI may have my knowledge, but it can’t have my heart.

I marveled at the savvy of this DiFi AI. She seemed to make the case for human emotions—without actually answering the question. And she didn’t count herself out of the race, either.

Could the human Feinstein do any better?