Democrats wanted an agreement on the use of artificial intelligence. Didn’t go anywhere

WASHINGTON – The Democratic National Committee was watching earlier this year as campaigns across the country experimented with artificial intelligence. So the organization approached a handful of influential party campaign committees with a request: sign guidelines that committed them to using technology in a “responsible” way.

The draft agreement, a copy of which was obtained by the Associated Press, was not full of revolutionary ideas. Asked campaigns to verify work through AI tools, guard against bias and avoid the use of AI to create misleading content.

“Our goal is to use this new technology effectively and ethically, and in a way that promotes – rather than undermines – the values ​​we espouse in our campaigns,” the draft said.

The plan came to nothing.

Rather than promoting agreement, the guidelines have sparked a debate about the value of such commitments, especially those governing rapidly evolving technology. Among the concerns expressed by Democratic campaign organizations: Such a commitment could harm their ability to implement AI and could alienate donors with ties to the AI ​​industry. Some committee staffers were also angry that the DNC gave them just a few days to agree to the guidelines.

The proposal’s disappearance highlighted internal divisions over campaign tactics and the party’s position. uncertainty about how to best utilize AI amid warnings from experts that technology is fueling the proliferation of misinformation.

Hannah Muldavin, a senior spokeswoman for the Democratic National Committee, said the group is not giving up on finding consensus.

The DNC, she said, “will continue to collaborate with our sister committees to discuss ideas and issues important to Democratic campaigns and American voters, including AI.”

“It is not uncommon for ideas and plans to change, especially in the middle of a busy election year, and any documents on this matter reflect early and ongoing conversations,” Muldavin said, adding that “the DNC and our partners take the opportunities and challenges presented by AI.”

The dispute arises as campaigns increasingly use artificial intelligence – computer systems, software or processes that emulate aspects of human work and cognition – to optimize workloads. This includes using great language models to write fundraising emails, send text messages to supporters, and build chatbots to answer voter questions.

This trend is expected to continue as the November general election approaches, with campaigns turning to supercharged generative AI tools to create text and images, as well as cloning human voices and creating videos at lightning speed.

The Republican National Committee used AI-generated images in a television commercial last year predicting a dystopian future under President Joe Biden.

Much of this adoption, however, has been overshadowed by concerns about how campaigns could use artificial intelligence in ways to mislead voters. Experts have warned that AI has become so powerful that it has made it easier to generate “deeply fake” videos, audio snippets and other media targeting opposing candidates. Some states have approved legislation regulating how generative artificial intelligence can be used. But Congress has so far failed to pass any bills regulating artificial intelligence at the federal level.

In the absence of regulation, the DNC searched for a set of guidelines it could point to as proof that the party was taking the threat and promise of AI seriously. The proposal was sent in March to the five Democratic campaign committees seeking to elect candidates for positions in the House, Senate, governor, state legislature and state attorneys general, according to the draft agreement.

The goal was to get each committee to agree on a list of AI protections, and the DNC proposed issuing a joint statement proclaiming that such guidelines would ensure campaigns could use “the tools they need to prevent the spread of misinformation and disinformation, while also empowering campaigns to act safely.” , responsibly use generative AI to engage more Americans in our democracy.”

The Democratic committee expected the statement to be signed by President Jaime Harrison and the leaders of the other organizations.

Democratic operatives said the proposal was received with a bang. Some senior committee leaders feared the deal could have unforeseen consequences, perhaps restricting how campaigns use AI, according to several Democratic operatives familiar with the disclosure.

And it could send the wrong message to technology companies and AI executives, many of whom help fill campaign coffers during election years.

Some of the Democratic Party’s most prolific donors are big tech entrepreneurs and AI evangelists, including Sam Altman, CEO of OpenAI, and Eric Schmidt, former CEO of Google.

Altman has donated more than $200,000 to Biden’s campaign and his Democratic-aligned joint fundraising committee since early last year, according to Federal Election Commission data, and Schmidt’s contributions to those groups have exceeded $ 500 thousand in the same period.

Two other AI proponents, Facebook co-founder Dustin Moskovitz and LinkedIn co-founder Reid Hoffman, have donated more than $900,000 to Biden’s joint fundraising committee this cycle, according to the same data.

The DNC’s plan caught committees off guard because it came with little explanation other than a desire to get each committee to agree on the list of best practices within a few days, said several Democratic operatives who spoke on condition of anonymity because they weren’t. authorized to discuss the matter. Aides to the Democratic Congressional Campaign and Democratic Senate Campaign committees said they felt rushed by a DNC timeline that urged them to sign quickly.

Representatives for the Democratic Attorneys General Association did not respond to The Associated Press’ request for comment. Spokespeople for the Democratic Governors Association and the Democratic Legislative Campaign Committee declined to comment.

The Republican National Committee did not respond to questions about its AI guidelines. The Biden campaign also declined to comment when asked about the DNC effort.

The four-page agreement — “Guidelines on the Responsible Use of Generative AI in Campaigns” — covered everything from ensuring that artificial intelligence systems could not be trusted without a human verifying their work to notifying voters when they were interacting with AI-generated content or systems.

“As the explosive rise of generative AI transforms every corner of public life – including political campaigns – it is more important than ever that we limit this new technology’s potential threat to voters’ rights and instead take advantage of it. it to build innovative, efficient campaigns and a stronger force. , a more inclusive democracy”, the proposal said.

The guidelines were divided into five sections that included headings such as “Providing Humane Alternatives, Consideration, and Recourse” and “Providing Notice and Explanation.” The proposed rules would require committees to ensure that “a real person should be responsible for approving AI-generated content and how, where, and to whom it is deployed.”

The directive outlined how “users should always be aware when interacting with an AI bot” and stressed that any images or videos created by AI “should be flagged” as such. And he emphasized that campaigns should use AI to help employees, not replace them.

“Campaigns are a human-driven and human-motivated business,” the agreement said. “Use the efficiency gains to educate more voters and focus more on quality control and sustainability.”

It also urged campaigns not to use “generative AI to create misleading content.” Period.”

___

This story is part of an Associated Press series, “The AI ​​Campaign,” exploring the influence of artificial intelligence on the 2024 election cycle.

___

The Associated Press receives financial assistance from the Omidyar Network to support coverage of artificial intelligence and its impact on society. AP is solely responsible for all content. Find APs standards for working with philanthropies, a list of supporters and areas of coverage funded in AP.org

____

The Associated Press and OpenAI have a licensing and technology agreement allowing OpenAI access to part of the AP’s text files.