🎸 Happy Tuesday, and happy birthday to the late Tom Petty.

☁️ Today's weather: Mostly cloudy with highs in the mid 60s, lows around 50.

🏡 Support coverage of the community you call home. Become a member today.

🏀 Situational awareness: The Valkyries fell 79-65 to the Atlanta Dream last night at the Chase Center, evening the WNBA Finals at 1-1 after this past weekend's thrilling Game 1 victory.

Today’s newsletter is 636 words — a 2.5-minute read.

San Francisco's congressional race offered an early glimpse of what happens when political campaigns use AI against their opponents, and it's playing out across the nation.

The big picture: State Sen. Scott Wiener's AI chatbot targeting Supervisor Connie Chan sparked a political firestorm this summer, and campaigns are ramping up their use of AI to tar their opponents as Election Day nears.

• They're putting words in opponents' mouths and creating realistic-looking scenes that never happened. And election laws largely haven't kept up.

Driving the news: More than $137 million has been spent on 220 AI-generated political ads through Oct. 11, mostly by Republicans, according to the Wesleyan Media Project.

• Only 27% of the identified ads disclosed their use of AI, according to the project.

Zoom in: Wiener's campaign launched a parody chatbot in August designed to mock Chan's positions on housing, transit and other issues.

• Instead, the bot generated responses about Chan's accent and citizenship that critics, including former House Speaker Nancy Pelosi, condemned as racist.
• Wiener initially defended the chatbot as satire before taking it down along with billboards promoting it.

Elsewhere, an Axios analysis found numerous political ads partly or fully made with generative AI, sometimes without disclosure.

• In Georgia, a Republican-aligned super PAC used AI to depict Democratic Sen. Jon Ossoff receiving beauty treatments.
• In Pennsylvania, a Republican attack ad featured an AI-generated version of Democratic Gov. Josh Shapiro making statements about taxes, abortion and data centers.

The big picture: Four states prohibit certain uses of AI-generated political deepfakes, while 27 others require some form of disclosure,.

• California has enacted restrictions, but courts have blocked parts of its AI election laws. No federal law specifically regulates AI-generated political ads.

The bottom line: Wiener's chatbot wasn't a traditional deepfake ad, but the controversy illustrated the same underlying risk: AI can make it harder for voters to separate a candidate's actual positions from an opponent's manufactured version.

Go deeper

💘 Dating in San Francisco has become an expensive, repetitive affair, with tech workers sporting nearly identical dating profiles and some singles shelling out thousands of dollars for matchmaking services, dating coaches and personal stylists. (SFGATE)

An ICE agent drew a gun on an apparently unarmed man in Soma yesterday, despite the man repeatedly telling agents he was a U.S. citizen, per security footage. (SF Standard)

🌊 The Bay Area could see its highest king tides in 18 yearsthis winter as a strong El Niño, warmer ocean temperatures and the moon's orbital cycle combine to push water levels higher. (CBS News)

A man stabbed aboard Muni's 14-Mission bus while intervening to protect female passengers from alleged sexual battery says he has no regrets about stepping in, even after surviving a horrific attack. (SF Standard)

🏃 About 15,000 runners turned out for Diplo's Run Club 5K in San Francisco over the weekend. (SFist)

👽 Looks like someone had a rough night.

• I spotted this little green alien sprawled out on the floor over the weekend at Pop's.

🛸 My thought bubble: Was this an extraterrestrial invasion gone wrong? A Halloween decoration that didn't survive the night? Or just another weekend casualty?

• Whatever happened, I have a feeling this wasn't the night out that our intergalactic visitor had planned.

😍 Nadia is so happy that she found a new pottery studio!

This newsletter was edited by Geoff Ziezulewicz.