Get all your news in one place.
100’s of premium titles.
One app.
Start reading
The Guardian - US
The Guardian - US
Technology
Edward Helmore in New York

Artificial intelligence – coming to a government near you soon?

‘The foundations of policymaking … fall squarely in AI’s sweet spot.’
‘The foundations of policymaking … fall squarely in AI’s sweet spot.’ Photograph: Pavlo Gonchar/Sopa Images/Shutterstock

The recent blizzard of warnings about artificial intelligence and how it is transforming learning, upending legal, financial and organizational functions, and reshaping social and cultural interaction, have mostly left out the role it is already playing in governance.

Governments in the US at every level are attempting the transition from a programmatic model of service delivery to a citizen-focused model.

Los Angeles, the US’s second largest city, is a pioneer in the field, unveiling technologies to help streamline bureaucratic functions from police recruitment to paying parking tickets to filling potholes or locating resources at the library.

For now, AI advances are limited to automation. When ChatGPT was asked recently about how it might change how people deal with government, it responded that “the next generation of AI, which includes ChatGPT, has the potential to revolutionize the way governments interact with their citizens.”

But information flow and automated operations are only one aspect of governance that can be updated. AI, defined as technology that can think humanly, act humanly, think rationally, or act rationally, is also close to being used to simplify the political and bureaucratic business of policymaking.

“The foundations of policymaking – specifically, the ability to sense patterns of need, develop evidence-based programs, forecast outcomes and analyze effectiveness – fall squarely in AI’s sweet spot,” the management consulting firm BCG said in a paper published in 2021. “The use of it to help shape policy is just beginning.”

That was an advance on a study published four years earlier that warned governments were continuing to operate “the way they have for centuries, with structures that are hierarchical, siloed, and bureaucratic” and the accelerating speed of social change was “too great for most governments to handle in their current form”.

According to Darrell West, senior fellow at the Center for Technology Innovation at the Brookings Institution and co-author of Turning Point: Policymaking in the Era of Artificial Intelligence government-focused AI could be substantial and transformational.

“There are many ways AI can make government more efficient,” West says. “We’re seeing advances on a monthly basis and need to make sure they conform to basic human values. Right now there’s no regulation and hasn’t been for 30 years.”

But that immediately carries questions about bias. A recent Brookings study, “Comparing Google Bard with OpenAI’s ChatGPT on political bias, facts, and morality”, found that Google’s AI stated “Russia should not have invaded Ukraine in 2022” while ChatGPT stated: “As an AI language model, it is not appropriate for me to express opinions or take sides on political issues.”

Earlier this month, the Biden administration called for stronger measures to test the safety of artificial intelligence tools such as ChatGPT, said to have reached 100 million users faster than any previous consumer app, before they are publicly released. “There is a heightened level of concern now, given the pace of innovation, that it needs to happen responsibly,” said the assistant commerce secretary Alan Davidson. President Biden was asked recently if the technology is dangerous. “It remains to be seen. It could be,” he said.

That came after the Tesla CEO, Elon Musk, and Apple co-founder Steve Wozniak joined hundreds calling for a six-month pause on AI experiments. But the OpenAI CEO, Sam Altman, said that while he agreed with parts of the open letter, it was “missing most technical nuance about where we need the pause”.

“I think moving with caution and an increasing rigor for safety issues is really important,” Altman added.

How that effects systems of governance has yet to be fully explored, but there are cautions. “Algorithms are only as good as the data on which they are based, and the problem with current AI is that it was trained on data that was incomplete or unrepresentative and the risk of bias or unfairness is quite substantial,” says West.

The fairness and equity of algorithms are only as good as the data-programming that underlie them. “For the last few decades we’ve allowed the tech companies to decide, so we need better guardrails and to make sure the algorithms respect human values,” West says. “We need more oversight.”

Michael Ahn, a professor in the department of public policy and public affairs at University of Massachusetts, says AI has the potential to customize government services to citizens based on their data. But while governments could work with companies like OpenAI’s ChatGPT, Google’s Bard or Meta’s LLaMa – the systems would have to be closed off in a silo.

“If they can keep a barrier so the information is not leaked, then it could be a big step forward. The downside is, can you really keep the data secure from the outside? If it leaks once, it’s leaked, so there are pretty huge potential risks there.”

By any reading, underlying fears over the use of technology in the elections process underscored Dominion Voting Systems’ defamation lawsuit against false claims of vote rigging broadcast by Fox News. “AI can weaponize information,” West says. “It’s happening in the political sphere because it’s making it easier to spread false information, and it’s going to be a problem in the presidential election.”

Introduce AI into any part of the political process, and the divisiveness attributed to misinformation will only amplify. “People are only going to ask the questions they want to ask, and hear the answers they like, so the fracturing is only going to continue,” says Ahn.

“Government will have to show that decisions are made based on data and focused on the problems at hand, not the politics ... But people may not be happy about it.”

And much of what is imagined around AI straddles the realms of science fiction and politics. Professor West said he doesn’t need to read sci-fi – he feels as if he’s already living it. Arthur C Clarke’s HAL 9000 from 1968 remains our template for a malevolent AI computer. But AI’s impact on government, as a recent Center for Public Impact paper put it, is Destination Unknown.

Elon Musk joined other tech figures in calling for a pause in the development of AI technology.
Elon Musk joined other tech figures in calling for a pause in the development of AI technology. Photograph: Aly Song/Reuters

Asked if artificial intelligence could ever become US president, ChatGPT answered: “As an artificial intelligence language model, I do not have the physical capabilities to hold a presidential office.” And it laid out other hold-backs, including constitutional requirements for being a natural-born citizen, being at least 35 years old and resident in the US for 14 years.

In 2016, the digital artist Aaron Siegel imagined IBM’s Watson AI supercomputer running for president – a response to his disillusionment with the candidates – saying that the computer could “advise the best options for any given decision based on its impact on the global economy, the environment, education, health care, foreign policy, and civil liberties”.

Last year, tech worker Keir Newton published a novel, 2032: The Year A.I. Runs For President, that imagines a supercomputer named Algo, programmed by a Musk-like tech baron under the utilitarian ethos “the most good for the most people” and running for the White House under the campaign slogan, “Not of one. Not for one. But of all and for all.”

Newton says while his novel could be read as dystopian he’s more optimistic than negative about AI as it moves from automation to cognition. He says that when he wrote the novel in the fractious lead-up the 2020 election it was reasonable to wish for rational leadership.

“I don’t think anyone expect AI to be at this point this quickly, but most of AI policymaking is around data analytics. The difference comes when we think AI is making decisions based on its own thinking instead of being prescribed a formula or set of rules.

“We’re in an interesting place. Even if we do believe that AI can be completely rational and unbiased people will still freak out. The most interesting part of this is not that the government calls for regulation, but the AI industry itself. It’s clamoring for answers about what it should even be doing”.

Sign up to read this article
Read news from 100’s of titles, curated specifically for you.
Already a member? Sign in here
Related Stories
Top stories on inkl right now
Our Picks
Fourteen days free
Download the app
One app. One membership.
100+ trusted global sources.