WASHINGTON (AP) — As the Democratic National Committee watched campaigns across the country experiment with artificial intelligence earlier this year, it asked several influential party campaign committees to sign guidelines committing to using the technology in a “responsible” way.
The draft agreement, obtained by The Associated Press, is hardly packed with revolutionary ideas: It calls for campaigns to check the work of AI tools, protect against bias and avoid using AI. Creating misleading content.
“Our goal is to use this new technology effectively and ethically, and in a way that advances rather than undermines the values we campaign for,” the draft states.
That plan went nowhere.
Rather than spurring consensus, the guidelines sparked debate about the value of pledges to govern a technology that is particularly rapidly evolving. Among the concerns expressed by Democratic campaign organizations was that such pledges could hinder the ability to adopt AI and alienate donors with ties to the AI industry. Some committee officials were outraged that the DNC gave them only a few days to agree to the guidelines.
The failure of this proposal was due to election strategy and party Uncertainty about how to best utilize AI Experts warn that the technology is accelerating the spread of disinformation.
Hannah Maldavin, a senior spokeswoman for the Democratic National Committee, said the committee hasn't given up on trying to reach an agreement.
He said the DNC would “continue to work with our sister committees to discuss ideas and issues important to Democratic campaigns and the American electorate, including AI.”
“It is not uncommon for ideas and plans to change, especially during a busy election period, and the documents on this subject reflect early and ongoing discussions,” Mardavin said, adding that “the DNC and our partners take seriously the opportunities and challenges that AI presents.”
The controversy comes as campaigns increasingly turn to artificial intelligence — computer systems, software and processes that mimic aspects of human tasks or cognition — to optimize their workloads, including using large-scale 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 using super-sophisticated generative AI tools to create text and images, replicate human voices, and create videos at lightning speed.
Republican National Committee Uses AI-generated images A television ad last year predicted a dystopian future under President Joe Biden.
But much of its adoption has been overshadowed by concerns that campaigns could use artificial intelligence to deceive voters. Experts have warned that AI has become so powerful that it could easily create “deep fake” videos, audio snippets and other media targeting opponents. Some states are allowing the use of artificial intelligence to target candidates. The bill was passed Regulating how generative artificial intelligence is used. However, Congress has so far failed to pass legislation to regulate artificial intelligence at the federal level.
In the absence of regulations, the DNC called for a set of guidelines the party could show as evidence it takes the threats and possibilities of AI seriously. According to the draft agreement, the DNC sent the proposal in March to five Democratic campaign committees seeking to select candidates for House, Senate, governor, state legislature and state attorney general.
The goal would be for each committee to agree to a set of AI guidelines, and the Democratic National Committee proposed issuing a joint statement declaring that such guidelines would ensure campaigns “use the tools necessary to prevent the spread of misinformation and disinformation and that campaigns use generative AI safely and responsibly to empower more Americans to participate in our democracy.”
The Democratic committee expected the statement to be signed by Chairman Jaime Harrison and other organization leaders.
The proposal came as a surprise to some committee leaders, who worried the deal could have unintended consequences and limit the use of AI in political campaigns, according to Democratic officials familiar with the effort.
It also might send the wrong message to tech companies and executives working on AI, many of whom are helping replenish campaign coffers in an election year.
Among the Democratic Party's biggest donors are top tech entrepreneurs and AI evangelists, including OpenAI CEO Sam Altman and former Google CEO Eric Schmidt.
According to Federal Election Commission data, Altman has donated more than $200,000 to the Biden campaign and its Joint Democratic Fundraising Committee since the beginning of last year, while Schmidt has donated more than $500,000 in the same period.
Two other AI advocates, Facebook co-founder Dustin Moskovitz and LinkedIn co-founder Reid Hoffman, also donated more than $900,000 to Biden's joint fundraising committee during the election, according to the same data.
The DNC's plan took the committees by surprise, with little explanation given beyond a desire to have committees agree on a list of best practices within a few days, said several Democratic officials who spoke on condition of anonymity because they were not authorized to discuss the matter. Aides to the Democratic House Campaign Committee and Democratic Senatorial Campaign Committee said they felt rushed by the DNC's timeline, which urged them to hurry the signatures.
A representative for the Democratic State Attorneys General Association did not respond to AP's request for comment, and spokesmen 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, and the Biden campaign declined to comment when asked about the DNC's efforts.
The four-page agreement, “Guidelines for the Responsible Use of Generative AI in Political Campaigns,” covers everything from ensuring artificial intelligence systems are not trusted without human checks to informing voters when interacting with AI-generated content and systems.
“As the explosion of generative AI transforms all aspects of public life, including political campaigns, it is more important than ever to limit the potential threats this new technology poses to voter rights and instead leverage it to build innovative, efficient campaigns and a stronger, more inclusive democracy,” the proposal reads.
The guidelines are divided into five sections, with titles such as “Providing Human Alternatives, Consideration, and Fallbacks” and “Providing Notice and Explanation.” The proposed rule called on the Commission to ensure that “real humans are responsible for approving AI-generated content and should be accountable for how, where, and to whom that content is deployed.”
The directive said “users should always exercise caution when interacting with AI bots” and stressed that images and videos created by AI “must be flagged as such.” It also stressed that AI should be used to assist campaigns, not replace them.
“Campaigns are human-driven and human-motivated businesses,” the agreement states. “We will use increased efficiencies to educate more voters and place an increased emphasis on quality control and sustainability.”
It also urged campaigns “not to use generative AI to create misleading content.”
___
This story is part of “The AI Campaign,” an Associated Press series exploring the impact of artificial intelligence in the 2024 election cycle.
___
The Associated Press receives financial support from Omidyar Network to support its coverage of artificial intelligence and its impact on society. The AP is solely responsible for all content. standard To work with a charity, see our list of supporters and funding areas. AP.org
____
The Associated Press and OpenAI License and Technology Agreements OpenAI will be providing access to portions of the AP's text archive.
