Most Americans want tech companies to commit to AI safeguards

Washington, DC, July 26, 2023 – Americans mostly support tech companies making a wide range of commitments to developing artificial intelligence (AI) responsibly. These readings are generally higher among Americans who have used AI tools recently, a group that is also the most optimistic about the future of AI. Meanwhile, an increasing number of Americans believe that AI will fundamentally change society as opposed to being another piece of technology among many.

Detailed findings

1. Fifteen percent of Americans say they have used an artificial intelligence (AI) chat or text program in the last month. Americans with a Bachelor’s (20%) or Master’s (20%) degree are more likely to say they have used AI in the last month than Americans with High school or equivalent (12%) or those with an Associate degree or some college (13%).
2. Half of all Americans think AI will fundamentally change American society over the next few years (51%), representing a 15-point rise compared to April 2023. In contrast, fewer think AI will simply be another piece of technology among many (41% in July, 55% in April).
   1. Enthusiasm with AI is higher among Americans who have used it in the last month. Two-thirds (68%) of these Americans think AI will fundamentally change society, as opposed to being another piece of technology (25%) or not have much of a real impact (7%).
3. The majority of Americans and AI users say they would support AI companies committing to adopting safety measures to ensure AI is developed responsibly. These include commitments to internal and external security testing (77%), investments in cybersecurity to protect model weights (72%), and the use of third parties to discover and report vulnerabilities (66%). Support for these measures is even more popular among AI users (83%, 81%, 83%, respectively).
   1. Most Americans (76%) say they support companies developing systems that would allow users to know when content was created by AI, including 80% of AI users.
   2. Americans also say they support AI companies committing to conduct research on the risks AI could pose to society (70%) as well as committing to developing AI systems to address society’s biggest challenges (64%).
4. Only a minority (32%) are familiar with news that the White House met with leaders of top AI companies, including Google, Meta, and OpenAI, to agree to voluntary commitments to develop AI technology safely, securely, and transparently. In contrast, AI users are over two times more likely to say they are familiar (63%) with this news compared to non-AI users (26%).
5. Americans are split when it comes to trusting AI companies to develop AI responsibly. Just under half of all Americans (48%) say they agree that they trust tech companies to uphold their commitment to developing AI responsibly, while 42% disagree. Trust of tech companies’ development of AI is higher among AI users (67%).
   1. The majority (71%) of Americans say Congress should pass a law turning these voluntary commitments into legal requirements, including 80% of AI users.
   2. A majority of Americans (69%) and AI users (80%) say they think that following these commitments to develop AI responsibly makes a positive future with AI more likely.

About the Study

These are the findings of an Ipsos poll was conducted July 24, 2023. For this survey, a sample of 1,000 adults age 18 or older from the continental U.S., Alaska, and Hawaii was interviewed online in English.

The sample was randomly drawn from Ipsos’ online panel, partner online panel sources, and “river” sampling and does not rely on a population frame in the traditional sense. Ipsos uses fixed sample targets, unique to each study, in drawing a sample. After a sample has been obtained from the Ipsos panel, Ipsos calibrates respondent characteristics to be representative of the U.S. Population using standard procedures such as raking-ratio adjustments. The source of these population targets is U.S. Census 2019 American Community Survey data. The sample drawn for this study reflects fixed sample targets on demographics. Post-hoc weights were made to the population characteristics on gender, age, race/ethnicity, region, education, political party affiliation and recent AI chat use. AI chat usage data was based on a recent Ipsos KnowledgePanel survey.  

Statistical margins of error are not applicable to online non-probability polls. All sample surveys and polls may be subject to other sources of error, including, but not limited to coverage error and measurement error. Where figures do not sum to 100, this is due to the effects of rounding. The precision of Ipsos online polls is measured using a credibility interval. In this case, the poll has a credibility interval of plus or minus 3.8 percentage points for all respondents. Ipsos calculates a design effect (DEFF) for each study based on the variation of the weights, following the formula of Kish (1965). This study had a credibility interval adjusted for design effect of the following (n=1,000, DEFF=1.5, adjusted Confidence Interval=+/-5.3 percentage points).

For more insights from this study, please check out our detailed paper or contact:

Chris Jackson
Senior Vice President, US
Public Affairs
+1 202 420-2025
[email protected]

About Ipsos

Ipsos is one of the largest market research and polling companies globally, operating in 90 markets and employing over 18,000 people.

Our passionately curious research professionals, analysts and scientists have built unique multi-specialist capabilities that provide true understanding and powerful insights into the actions, opinions and motivations of citizens, consumers, patients, customers or employees. Our 75 solutions are based on primary data from our surveys, social media monitoring, and qualitative or observational techniques.

Our tagline "Game Changers" sums up our ambition to help our 5,000 customers move confidently through a rapidly changing world.

Founded in France in 1975, Ipsos has been listed on the Euronext Paris since July 1, 1999. The company is part of the SBF 120 and Mid-60 indices and is eligible for the Deferred Settlement Service (SRD).ISIN code FR0000073298, Reuters ISOS.PA, Bloomberg IPS:FP www.ipsos.com