Six in Ten (61%) Respondents Across 26 Countries Oppose the Use of Lethal Autonomous Weapons Systems

Opposition to Fully Autonomous Weapons Has Increased Since 2017, Up From 56%

22 January 2019

The author(s)

Chris Deeney Senior Vice President, U.S. Ipsos Observer

Get in touch

Washington, DC, January 22, 2019 — According to a recent online survey conducted by Ipsos on behalf of Human Rights Watch for the Campaign to Stop Killer Robots, sixty one percent of adults across 26 countries say that they oppose the use of lethal autonomous weapons systems, also known as fully autonomous weapons. On the other hand, 22 percent support such use and 17 percent say that they are not sure. In a similar study conducted by Ipsos in January 2017, 56 percent were opposed, 24 percent not opposed, and 19 percent unsure.

Support for fully autonomous weapons is strongest in India (50%) and Israel (41%). The strongest opposition is in Turkey (78%), South Korea (74%), and Hungary (74%).

Among those who are opposed, 66% say that they feel this way because they believe lethal autonomous weapons systems cross a moral line as machines should not be allowed to kill. More than half (54%) of those who are opposed also feel this way because weapons are “unaccountable.”

The United Nations is reviewing the strategic, legal and moral implications of lethal autonomous weapons systems. These weapons systems would be capable of independently selecting targets and attacking those targets without human intervention. They are thus different than current day "drones" where humans select and attack targets. How do you feel about the use of such lethal autonomous weapons systems in war?

	Base	Strongly/ Somewhat support	Somewhat/ Strongly oppose
Total	18,795	22%	61%
India	576	50%	37%
Israel	500	41%	41%
Brazil	1,003	36%	46%
China	1,001	32%	60%
Mexico	502	28%	64%
Peru	501	27%	65%
Argentina	502	26%	64%
South Africa	591	26%	59%
United States	1,000	24%	52%
Poland	501	22%	62%
Great Britain	1,001	21%	54%
Italy	1,001	20%	58%
Colombia	501	20%	73%
Spain	1,001	19%	65%
Russia	500	18%	59%
Belgium	501	17%	63%
France	1,002	16%	59%
South Korea	501	15%	74%
Sweden	502	15%	71%
Australia	1,000	15%	59%
Canada	1,002	15%	60%
Germany	1,002	14%	72%
Japan	1,102	14%	48%
Hungary	500	13%	74%
Turkey	500	12%	78%
Netherlands	502	12%	68%

If you oppose the use of lethal autonomous weapons systems, which of the following concerns you?

	Total	Argentina	Belgium	Mexico	Poland	Russia	South Africa	South Korea	Sweden	Turkey	Hungary	Australia	Brazil	Canada
Base: Those opposed	11,429	322	333	321	324	311	364	360	374	372	367	606	489	652
They'd be illegal	21%	22%	18%	25%	19%	21%	25%	12%	22%	26%	17%	19%	18%	18%
They'd be unaccountable	54%	76%	48%	75%	51%	60%	51%	44%	53%	56%	52%	65%	32%	62%
They'd cross a moral line because machines should not be allowed to kill	66%	58%	69%	55%	68%	70%	74%	63%	73%	54%	81%	68%	54%	67%
They'd be too expensive	10%	6%	17%	9%	10%	4%	18%	7%	6%	5%	11%	12%	11%	16%
They'd be subject to technical failures	45%	48%	47%	52%	35%	33%	46%	47%	56%	17%	47%	65%	54%	59%
Something else	9%	7%	8%	6%	9%	4%	9%	2%	10%	21%	10%	10%	10%	8%
I don't know	4%	1%	5%	1%	8%	4%	2%	6%	4%	5%	3%	3%	8%	5%

	Total	China	France	Germany	Great Britain	India	Italy	Japan	Spain	United States	Peru	Israel	Netherlands	Colombia
Base: Those opposed	11,429	603	596	743	549	214	585	552	646	525	323	205	331	362
They'd be illegal	21%	18%	40%	19%	19%	20%	33%	19%	15%	19%	15%	31%	13%	27%
They'd be unaccountable	48%	62%	64%	35%	48%	72%	41%	45%	45%	36%	66%	77%	48%	74%
They'd cross a moral line because machines should not be allowed to kill	66%	67%	85%	70%	77%	66%	68%	65%	59%	70%	62%	50%	65%	59%
They'd be too expensive	10%	16%	6%	10%	7%	13%	15%	10%	9%	9%	18%	8%	11%	9%
They'd be subject to technical failures	45%	59%	26%	56%	60%	60%	43%	32%	31%	19%	72%	41%	39%	44%
Something else	9%	8%	4%	7%	7%	7%	11%	10%	7%	10%	12%	4%	15%	8%
I don't know	4%	5%	1%	5%	5%	3%	2%	5%	7%	7%	2%	2%	7%	2%

About the Study

The findings come from surveys conducted between November 26 to December 7, 2018 on the Ipsos Global Advisor platform using the Ipsos Online Panel system with 18,795 adults across 26 countries: Argentina, Australia, Belgium, Brazil, Canada, China, Colombia, France, Germany, Great Britain, Hungary, India, Israel, Italy, Japan, Mexico, Netherlands, Peru, Poland, Russia, South Africa, South Korea, Spain, Sweden, Turkey, and the U.S.

The sample size per country in each survey is approximately N=1,000 for Australia, Brazil, Canada, China, France, Germany, Great

Britain, Italy, Japan, Spain, and the U.S. and approximately N=500 for Argentina, Belgium, Colombia, Hungary, India, Israel, Mexico, Netherlands, Peru, Poland, Russia, South Africa, South Korea, Sweden, and Turkey.

Weighting has been employed to balance demographics and ensure that the sample’s composition reflects that of the adult population according to the most recent country census data.

In 16 of the countries surveyed, internet penetration is sufficiently high to think of the samples as representative of the national population within the age ranges covered: Argentina, Australia, Belgium, Canada, France, Germany, Hungary, Italy, Japan, Netherlands, Poland, South Korea, Spain, Sweden, Great Britain, and the U.S. Brazil, Colombia, China, India, Israel, Mexico, Russia, Peru, South Africa, and Turkey have lower levels of internet penetration. Samples from those countries should not be considered fully nationally representative, but instead to represent a more affluent, connected population, representing an important and emerging middle class.

All survey respondents are aged 18 ‐ 64 in Canada and the U.S. and 16 ‐ 64 in all other countries. The precision of Ipsos online polls is measured using a credibility interval. In this case, the poll has a credibility interval of ±3.5 percentage points for a sample of 1,000 and of ±5.0 percentage points for a sample of 500.

Statistical margins of error are not applicable to online nonprobability sampling 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. 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=5.0 and n=500, DEFF=1.5, adjusted Confidence Interval=6.5).

For more information about conducting research intended for public release or Ipsos’ online polling methodology, please visit our Public Opinion Polling and Communication page where you can download our brochure, see our public release protocol, or contact us.

For more information on this news release, please contact:

Chris Deeney
Senior Vice President, U.S.
Ipsos Public Affairs
+1 312 526-4088
[email protected]

Download

press release

The author(s)

Chris Deeney Senior Vice President, U.S. Ipsos Observer

Society