There’s a famous saying about Statistics being the art of misleading people into believing that, if two individuals had two chickens available and one person ate both, what actually happened is that they ate one each.
I could not avoid thinking of that, when I came across the Oracle-commissioned "From Fear to Enthusiasm - AI is Winning Hearts and Minds in the Workplace" report, a couple of days ago.
Research firm Future Workplace surveyed 8,370 employees, managers and HR persons across 10 countries, clever PR followed and it all translated into headlines such as "64% of people trust a Robot more than their Manager", "Four out of five think robots are better than their managers" or "Chinese workers trust robot coworkers more than their managers".
But is this what the report says?
Here things become a little problematic.
Skeptical corporate employees at work looking at robot colleague with skepticism and jealousy while ... [+]
It is true that, according to the study, 64% of respondents would trust a robot more than their managers, but the difference in feedback between French, American and UK workers on one side and Eastern workers on another, is so huge that it’s almost as if you were looking at two different surveys.
The pro-AI message of the survey relies a lot on the fact that 90% of Indian, 88% of Chinese and 84% of Singaporean interviewees answered "yes" to the "would you trust a robot more than your manager" question. By contrast, only 52% of French, 55% of Brits and 57% of Americans agreed. Still a slight majority, but quite a different picture.
What to make of it? Does it mean that the latter are less forward-looking and ready to embrace innovation in general? Or that they are more conscious and wary of the consequences of automation in the workplace? We don’t know for certain.
What we know is that the study was administered online, "respondents were recruited through a number of mechanisms, via different sources", had passed "a double opt-in process and complete, on average, 300 profiling data points" and were "invited to take part via email and are provided with a small monetary incentive for doing so."
The report does indeed offer a possible explanation for these different outcomes: the countries in which respondents say AI adoption is stronger - India, China, but also Brazil - and the chances of experimenting with it are higher, are also the ones in there’s more optimism and excitement about the opportunities involved.
According to this line of reasoning, Western workers and managers are less bullish about AI because they are much more ignorant and less aware of its capabilities than Brazilian or Indian ones.
Another possible explanation is that they are more concerned than their Asian counterparts about the need for human interaction, security and privacy in the workplace, the main factors that according to the report are preventing companies from using AI.
Making sense of the report’s results gets even more complicated if you look at the size of the samples involved. As we have seen, the impact of China, India and Singapore on the overall AI acceptance score is very significant; surprisingly, though, the samples considered for those countries are very small, especially as a percentage of the respective populations.
A screenshot from the report, showing the number of respondents by country
I could not find exact figures, but, judging from a graph, for China, fewer than 1,000 people were interviewed. Therefore, as Dan Robitzki of Futurism also notes, it’s not clear how could they be representative of the 800 million people in the country’s workforce.
My University notions of Statistics are a bit rusty, unfortunately, but I wonder, based on the above, if it’s really possible to say that "nearly 90% of Chinese workers have more trust in robots than in their human managers."
Isn’t it a bit far-fetched?
