Top 20 artificial intelligence companies, 83 countries, 7-step pledge to ban election interference, joint response to fake news
Twenty artificial intelligence (AI) companies announced seven joint protocols for response to ‘election interference’ in elections in 83 countries, including the mandatory attachment of ‘easy identification labels’.
Adobe, Amazon, Anthropic, Google, Meta, Microsoft, OpenAI, TikTok, X, etc. 20 Several artificial intelligence technology companies announced in a press release on the 16th a voluntary pledge to prevent fraudulent artificial intelligence content from interfering with voting in this year's election.
The TECHNOLOGY INDUSTRY TO COMBAT DECEPTIVE USE OF AI IN 2024 ELECTIONS agreement announced at the Munich Security Conference included ‘a commitment to cooperate on artificial intelligence detection tools and other measures’, but did not enforce a ban on election-related artificial intelligence content.
The Technical Agreement to Prevent the Deceptive Use of AI in the 2024 Elections states that “signatories may “We set expectations on how the risks that arise will be managed in accordance with our own policies and practices related to the commitments of the agreement,” it said.
The agreement presented seven national goals, including ‘prevention’, ‘source attribution’, ‘detection’, ‘response protection’, ‘public education’, and ‘protection of democratic procedures’.
First, under ‘prevention’, “invest in research and deploy reasonable precautions to limit the risk of intentionally deceptive AI election content being created,” and under ‘source attribution’, “where appropriate and technically feasible, attach a source signal to identify the source of the content.” ” was enforced.
The ‘Detection’ phase will jointly respond to “attempts to detect fraudulent AI election content or authenticated content (methods such as reading cross-platform source signals)” and the ‘Response Protection’ phase will “respond to incidents related to the creation and dissemination of fraudulent AI election content.” “Providing a quick and balanced response” was guaranteed.
In ‘Evaluation’, we place emphasis on joint evaluation and collective verification, specifying “Evaluate the experience and results of dealing with fraudulent AI election content and conduct collective efforts to learn,” and to expand ‘public awareness,’ “Best practices in media literacy, especially fraudulent AI.” “Participate in concerted efforts to educate the public regarding content and how citizens can protect themselves from being manipulated or deceived by such content.”
The final phase, ‘Resilience,’ commits to “support efforts to develop and deliver defense tools and resources, such as AI literacy, other public programs, AI-based solutions (including open source tools), or contextual capabilities,” and “protect public debate; , defend the integrity of the democratic process, and build the resilience of society as a whole against the use of deceptive AI election content.”
Anthrophic separately announced on this day that its technology ‘will be prohibited from being applied to political campaigns or lobbying.’
In a blog post, Anthropic said it was building a chatbot called Claude that would “warn or suspend users who violate its rules” and “train it to automatically detect and block misinformation and influence operations.” “We are using tools that have been developed,” he said of additional measures.
OpenAI, the maker of the ChatGPT chatbot, previously took steps to prevent abuse of the tool in elections last month by partially banning its use to create chatbots that impersonate real people or organizations.
The New York Times said that artificial intelligence companies have begun setting limits on how they use AI, given that the development of innovative technologies is amplified in a year that coincides with major elections around the world. “In recent weeks, Google has launched its artificial intelligence chatbot Bard. It said it would "out of an abundance of caution" limit responses to certain election-related prompts.
The NYT said Meta, which owns Facebook and Instagram, will better label artificial intelligence-generated content on its platforms to make it easier for voters to discern which material is real and which is fake.
Synthesia, a startup whose AI video generator has been linked to disinformation campaigns, announced a ban on using its technology for "news-like content" that includes material that is false, polarizing, divisive or misleading. .
“The company has improved the systems it uses to detect misuse of its technology,” Alexandru Voica, director of corporate affairs and policy at Synthesia, told the NYT today.
Stability AI, a startup with an image generator tool, prohibits the use of its technology for illegal or unethical purposes, strives to block the creation of unsafe images, and “invisibly removes water from all images.” “Apply the mark.”
In a press release on December 19 last year, Google said, “We will require YouTube video creators and all election advertisers to disclose digitally altered or created content.”
Google “focuses on a principled and responsible approach when adopting generative AI products, including Search Generative Experience (SGE) and Bard, and prioritizes testing for safety risks, from cybersecurity vulnerabilities to misinformation and fairness.” “To ensure we get the most attention to these important topics in preparation for the 2024 elections, we will be limiting the types of election-related queries for which Bard and SGE will return responses.”
At least 83 elections are expected to be held around the world this year, making it the most election year in the next 24 years.
In recent weeks, people in Taiwan, Pakistan, and Indonesia have voted, and India, the world's largest democracy, is holding general elections in March, followed by South Korea's general elections.