Three-quarters of Americans are deeply concerned about the possible harms that could be caused by misuse of generative artificial intelligence (GAI) tools such as ChatGPT, according to a new survey from ADL (the Anti-Defamation League) released today.
While many are hopeful about the potential for GAI to improve their lives, the survey found that 75 percent of Americans are deeply concerned the technology could be “misused or abused” to incite hate and harassment. The survey also found that 75 percent of respondents believe the tools will produce biased content; and 70 percent believe GAI tools will make extremism, hate and antisemitism worse in America.
“If we’ve learned anything from other new technologies, we must protect against the potential risk for extreme harm from generative AI before it’s too late,” said Jonathan A. Greenblatt, ADL CEO. “We join with most Americans in being deeply concerned about the potential for these platforms to exacerbate already high levels of antisemitism and hate in our society, and the risk that they will be misused to spread misinformation and fuel extremism.”
The survey results were released in advance of a Senate hearing devoted to the issue, where OpenAI CEO Sam Altman and other officials are expected to be questioned about the potential risks posed by AI chatbots and other generative technologies.
Among the survey’s key findings:
- Americans overwhelmingly support interventions to mitigate the risks posed by generative AI. Nearly 90 percent believe that companies should take steps to prevent their tools from creating harmful content and not allow users to generate antisemitic or extremist images.
- Respondents overwhelmingly support government intervention, with 87 percent supporting congressional efforts to mandate transparency and privacy, and 81 percent saying creators of GAI tools “should be held responsible” if their tools are used for illegal hate, harassment or extremism.
- Eighty-four percent (84%) are concerned that GAI could be used for criminal purposes, like fraud or identity theft.
- Seventy-seven percent (77%) fear extremists will use the tools to radicalize more people.
- Seventy-five percent (75%) believe the tools will produce content that is biased against marginalized groups.
- Americans overwhelmingly believe civil society should have the ability to audit generative AI tools, with 85 percent agreeing that academics or civil society groups “should have access to review or audit the tools to make sure they are properly constructed.”
In a new blog, “Six Pressing Questions We Must Ask About Generative AI,” ADL urged policymakers and industry professionals to implement safeguards to prevent the technology from being weaponized to sow disinformation or harassment, or to fuel extremism.
“There’s no doubt that many exciting technological advancements are possible with the increased access to GAI,” said Yael Eisenstat, Vice President of the ADL Center for Technology & Society. “But this technology may be abused to further accelerate hate, harassment and extremism online. As lawmakers and industry leaders prioritize innovation, they must also address these challenges to prevent their misuse. We look forward to working with policymakers, industry leaders and researchers as they establish standards for GAI.”
The survey of 1,007 U.S. adults was fielded by the ADL Center for Technology & Society using Qualtrics, a nationally representative survey platform, between May 1-5, 2023.
Building on ADL’s century of experience building a world without hate, the ADL Center for Technology & Society serves as a resource to tech platforms and develops proactive solutions to fight hate both online and offline. CTS works at the intersection of technology and civil rights through education, research and advocacy. Last year ADL announced the REPAIR Plan, a comprehensive framework for platforms and policymakers to take meaningful action to decrease online hate and extremism.
