Cal is Program of public consultation. Lewitus is a research analyst at Voice of the People. Thomas is Vice President of Voice of the People and Director of Voice of the People Action.
A new House of Commons task force on artificial intelligence is considering how the government should address AI issues such as election-related deepfakes and algorithmic bias, but a new study calls for the federal government to regulate AI. It found a bipartisan majority in favor of granting broad powers. .
They support seven proposals currently under consideration in Congress and the executive branch to regulate AI-generated deepfakes and AI-generated decision-making that could cause harm. Voters also support an international treaty that would ban AI-controlled weapons and establish an international body to regulate large-scale AI projects.
The study was conducted by the Public Consultation Program at the University of Maryland School of Public Policy. As with all public consultation surveys, respondents will be provided with a detailed explanation and discussion of the pros and cons of each proposal to ensure that they fully understand their AI issues. Reviewed by experts on each side of the debate.
The enactment of new legislation against AI-generated deepfakes recorded overwhelming bipartisan support. Advances in AI have made it easier to create highly realistic images, video, and audio. All three proposals surveyed received support from more than 8 in 10 Republicans and Democrats. they would do this:
Sign up for the Fulcrum newsletter
- Ban the use of deepfakes in political campaign ads, as proposed by the Federal Election Commission (84 percent of national parties, 83 percent of Republicans, 86 percent of Democrats). For example, expressing something that the opposition did not say, or depicting an event that did not actually occur. percent).
- Prohibit the public distribution of pornographic deepfakes created without the consent of the person depicted, as proposed in the Intimate Image Deepfakes Prevention Act and the DEFIANCE Act. (National Party 86%, Republican Party 85%, Democratic Party 87%)
- Require all publicly shared deepfakes to be clearly labeled as such, as proposed in the AI Labeling Act, the AI Disclosure Act, and the Deepfake Accountability Act. (National Party 83%, Republican Party 83%, Democratic Party 85%)
A bipartisan majority also believes that decisions that can have a significant impact on people's lives, such as health care, banking, employment, criminal justice, and welfare, should be made in ways similar to how the FDA regulates drugs. We support three proposals to strictly regulate AI programs that perform There is evidence that some of these programs violate regulations, exhibit bias (based on race, gender, age, etc.), and cause serious harm to individuals.
More than 7 in 10 voters support proposals that:
- Require these AI programs to pass tests before they can be used. Testing assesses whether it may violate regulations, make biased decisions, or have security vulnerabilities. (National Party 81%, Republican Party 76%, Democratic Party 88%)
- Allow the government to audit the programs they use and require AI companies to fix any problems found (77% national, 74% Republican, 82% Democratic).
- Require AI companies to disclose to the government, upon request, information about how their decision-making AI was trained to assist with pre-testing or audits (72 percent nationally, 67 percent Republicans, 81 percent Democrats). ).
These proposals are based on the Algorithmic Liability Act and mirror the European Union's Artificial Intelligence Act regulations.
Creating a new federal agency for AI that would enforce regulations, oversee development, and provide policy guidance is supported by 74 percent (68 percent Republicans, 81 percent Democrats). This proposal is based on the Digital Platforms Commission Act.
On the international front, Americans also support the creation of an international regulatory body for large-scale AI, modeled on the International Atomic Energy Agency, as proposed by OpenAI, New York University professor Gary Marcus, and United Nations Secretary-General Antonio Guterres. A bipartisan majority (77%) supports creating an international body with the power to set standards for large-scale AI and monitor and inspect whether those standards are met (71% Republicans; Democratic Party (84%).
In the international arena, Americans also advocate for a treaty to ban the development of lethal autonomous weapons, which can use AI to attack targets without human control, as called for by the International Committee of the Red Cross. I support it. A campaign to stop killer robots. A bipartisan majority (81%) supports the United States actively working to create such treaties and the creation of an international body to enforce the ban (Republicans 77%, Democrats 85%).
Clearly, Americans are seriously concerned about the current and potential harms of AI. Majorities of Republicans, as well as Democrats, favor the creation of strong federal and international institutions to regulate AI and protect people from deepfakes, biased decision-making, and other potential harms from AI. There is.
When respondents evaluated the arguments for and against each of the proposals listed above, they found that arguments in favor of regulation were persuasive by majorities of both Republicans and Democrats. However, the majority also found a number of counter-persuasive arguments, including: Regulation will stifle innovation. Prohibitions violate freedom of expression. And international organizations can abuse their power.
Americans are wary of government regulation, but they are clearly wary of unrestrained development and use of AI.
The survey was conducted online from February 16 to 23, 2024 using a representative, non-probability sample of 3,610 registered voters provided by Precision Sample from its large online panel. Confidence intervals vary from +/- 1.4 to 1.8 percent.
Read the full report And that questionnaire.
From an article on your site
Related articles on the web
