With help from Derek Robertson

NEW YORK — On a Harlem street this summer, New Yorkers caught a glimpse of the future.

Strutting between a logjam of NYPD vehicles blocking an intersection was one of the NYPD’s newest recruits: a robotic canine called Digidog, emblazoned with the department’s blue and white colors and outfitted with a number of high-tech accessories.

The funds to purchase the cybernetic hound did not go through the standard budgeting process, which requires oversight and a vote from the New York City Council. Instead, police brass received cash directly from the federal government under something called the Equitable Sharing Program, which supplements the budgets of local police departments with money and property forfeited in the course of criminal investigations.

The multi-billion dollar initiative has helped law enforcement agencies pay overtime and arm themselves with equipment and sophisticated weaponry since the Reagan era. But the program is now entering a new phase as it provides access to a futuristic era of high-tech policing tools that have raised fresh questions about the balance between privacy and public safety, along with biases inherent in supposedly neutral algorithms.

Advances in artificial intelligence, surveillance and robotics are putting the stuff of yesteryear’s science fiction into the hands of an ever-growing list of municipalities from New York City to Topeka.

Privacy advocates are worried both about the Digidog — the devices can collect data and the NYPD has not thoroughly disclosed how that information would be used — and about high-tech policing tools in general.

“More departments are using more tools that can collect even more data for less money,” said Albert Fox Cahn, head of the New York City-based watchdog group Surveillance Technology Oversight Project. “I’m terrified about the idea that we’ll start seeing decades of work to collect massive databases about the public being paired with increasingly invasive AI models to try to determine who and who isn’t a threat.”

While recognizing that technology can sometimes be a helpful tool to fight crime, privacy advocates nevertheless worry about a lack of guardrails around the ethics of police departments using robots, facial recognition and increasingly broad local surveillance networks.

At the end of a press release announcing the purchase of two Digidogs, for instance, the NYPD sought to assuage a concern grimly indicative of this new era.

“Under the NYPD’s protocols, officers will never outfit a robot to carry a weapon and will never use one for surveillance of any kind,” the department wrote.

It turns out, that’s an important disclaimer. Companies like Ghost Robotics have already attached sniper rifles to quadruped robots. And in November, the San Francisco Board of Supervisors voted to give law enforcement robots the authority to use lethal force.

That proposal — which would have allowed police to place explosives on automatons in limited circumstances — was reversed after public outcry. But the board left the door open to reconsidering the initiative in the future.

Other technology seems to have biases baked into its foundation, with serious implications for communities of color. And vast amounts of biometric data, along with license plate readers that can pinpoint the location of a particular vehicle, are creating the capability for broad surveillance of the citizenry.

“In our country, the police should not be looking over your shoulder, literally or figuratively, unless they have an individualized suspicion that you are involved in wrongdoing,” Jay Stanley of the American Civil Liberties Union said in an interview. “They can’t just watch everybody all the time in case you commit a crime.”

Alongside the new concerns that come with each technological advancement, the money underwriting some of these products is also under increasing scrutiny.

Read the full story from POLITICO New York here.

United Kingdom Prime Minister Rishi Sunak might be running out of time to deliver on his big AI promises.

That’s according to a report this morning from POLITICO’s Tom Bristow, Gian Volpicelli and Laurie Clarke, on how allies are pressuring the U.K. to put its ambitious plans for a “major global summit” on AI into action. Private and public officials alike are unsure whether it will actually happen, with key details about the dates and attendees still up in the air.

Two anonymous embassy officials told them that they “haven’t been given much information yet,” “And we’re getting quite close to the date now,” and one diplomat said the U.K. clearly needs to “step up” its summit planning game. The reporters point out that responsibility for the details “is split between U.K. government teams in DSIT, the Foreign Office and No.10.”

One major looming question is whether China should be invited, as the democratic world largely and increasingly sees the country as an adversary bent on using AI tools for authoritarian surveillance. But many officials are arguing that not including the world’s second-biggest economic and tech superpower in a global AI summit would render such a thing pointless: Matt Sheehan, a fellow at the Carnegie Endowment for International Peace, said that “If you’re trying to solve the core problem of AI safety on a global level, you can’t not include China.” — Derek Robertson

The Federal Election Commission voted unanimously today to advance a petition for rulemaking around the use of AI-assisted deepfakes in elections.

The petition, brought by progressive nonprofit Public Citizen, urges the FEC to specifically include AI-generated material under its definition of “fraudulent misrepresentation” in political ads which is already prohibited. A 60-day public comment session will now open, after which the Commission will decide whether to make a rule.

Listen to today’s edition of the POLITICO Tech podcast, where Steven Overly speaks with Public Citizen’s president Robert Weissman about the dangers AI might pose, or is already posing, to elections, and the political considerations involved in specifically prohibiting its use. — Derek Robertson

• Computer scientists are using AI tools to fight sexual exploitation.
• Prompt injection is an increasingly popular, and simple, way to do cybercrime.
• Listen to the audio story of a man whose life is changed by an AI companion.
• The head of Alibaba warns that geopolitics are slowing down China’s AI push.
• Grimes waxes poetic on her AI fantasia.

Stay in touch with the whole team: Ben Schreckinger ([email protected]); Derek Robertson ([email protected]); Mohar Chatterjee ([email protected]); and Steve Heuser ([email protected]). Follow us @DigitalFuture on Twitter.

If you’ve had this newsletter forwarded to you, you can sign up and read our mission statement at the links provided.