Twenty technology companies, including Google, Meta, OpenAI, and We are committed to eliminating.
The corporate group is Technology Agreement to Combat AI Abuse in 2024 Elections The Munich Security Conference outlines a “voluntary framework of principles and actions” aimed at preventing, detecting, responding to, assessing and identifying sources of fraudulent AI election content.
Related article: Ethics of generative AI: How can you take advantage of this powerful technology?
It also includes efforts to raise public awareness about how to protect themselves from manipulation by such content, according to a joint statement issued by signatories of the technology agreement, which also includes TikTok, Amazon, IBM, Anthropic and Microsoft. .
The agreement commits the 20 organizations to eight mission statements, including seeking to use AI to detect and prevent the distribution of deceptive election content; This includes providing transparency to the public about how to deal with it. The two companies will work together to develop and implement tools to identify content, curb its spread, and track the origin of such content.
These efforts include developing methods and standards for classifiers and provenance, such as watermarks and signed metadata, and attaching machine-readable information to AI-generated content.
The Eight Promises apply as they relate to the services each company provides.
Related article: Election 2024: How AI will fool voters if we don’t do something now
The agreement covers content defined as “persuasive” AI-generated audio, video, and images, including the “appearance and appearance” of political candidates, election officials, and other key election stakeholders. “Deceptively disguise or alter voice, conduct, or otherwise promote an election.” Providing the public with fraudulent information about where, when, and how to vote.
“In 2024, more people will hold elections than in any other year in history, and more than 4 billion people in more than 40 countries will choose their leaders and representatives,” the technical agreement states. There is. “At the same time, the rapid development of AI is creating new opportunities and challenges for democratic processes. The whole of society can rely on the opportunities presented by AI and take new steps together to protect elections and electoral processes.” will have to step up in this exceptional year. ”
Also: We’re not ready for how generative AI will impact elections
The agreement sets expectations for how signatories will manage risks arising from deceptive AI election content created via public platforms and open infrastructure models, or distributed on social and publishing platforms. The purpose is that. These are consistent with the signatories’ own policies and practices.
Models and demonstrations intended for research purposes or primarily for corporate use are not covered by this agreement.
The signatories added that AI can be leveraged to help defenders counter bad actors and detect deceptive campaigns faster. AI tools can also significantly reduce overall defense costs, allowing even smaller organizations to implement robust protection.
“While we are committed to our role as a technology company, we recognize that the misuse of AI is not only a technical challenge, but also a political, social and ethical issue. We hope that other companies will similarly take action across society,” the signatories said. Said. “We affirm that protecting election integrity and public trust is a shared responsibility and common interest that transcends partisan interests and national borders.”
Also: Want to work in AI? How to pivot your career in 5 steps
Munich Security Council President Christoph Heusgen said the agreement was an “important step” in advancing electoral integrity and societal resilience. It would also help build “trusted technology practices,” he said.
Risks related to AI-powered misinformation on social cohesion will dominate this year, according to the 2024 Global Risks Report released by the World Economic Forum (WEF) last month. The report names misinformation and disinformation as a key global risk over the next two years, warning that its widespread use and the tools to spread it could undermine the new government’s legitimacy. are doing.