We’re not that positive about new technology. Typically, we treat emerging tech with a healthy dose of cynicism. Occasionally, we treat it with outright hostility. 

The chart below shows the number of people agreeing or disagreeing with the statement: “Generally, new technology has a positive impact on society.” We’ve asked this question in the Trajectory Optimism Index every month since February 2018, and the data below are three-month rolling averages.

 

We can’t be sure exactly what people are picturing when they think about ‘new technology’. Whatever we’re thinking of, we’re not that positive. In the first five months of the year, on average, fewer than 60% agree that new technology has a positive impact on society.

Paranoid about androids

Artificial Intelligence has been around for decades. In this sense, although current manifestations are new, the concept is very established. Since at least the 1950s, and the IBM 702 and the Dartmouth Conference, scientists, researchers and science fiction writers have imagined a future in which computer systems simulate human intelligence. For decades it manifested mostly in fiction – with audiences more likely to be exposed to AI at the cinema than in day-to-day life – but since 2022 more people have been able to use forms of AI in publicly available tools.

Firstly, some statements on AI.

Only a third of people think that either the public or private sector should make more use of AI and only 22% think it will make society fairer. Although people aren’t that worried about the impact of AI on their job (30%) more than twice the proportion are worried about the impact on society (61%). People are particularly concerned about deepfakes (62%). Election manifesto writers take note: 67% think the government needs to do more to regulate and control the development and use of the technology; only 9% would oppose that.

This is a far cry from the lukewarm reception that technology generally gets in our tracker. This is hostility, rather than a shrug. It’s worth noting that a sizeable proportion – usually around two-fifths – of the public either don’t know or don’t have much of an opinion on any of the statements. This partly reflects the relatively low salience of the issue, but also how little people even know about AI.

Less than half of the population have even heard the term ‘generative AI’. Only a third have heard the term ‘automation’ in the context of new technology. Despite what it might feel like on SubStack, LinkedIn or X/Twitter, the whole world is not talking about AI.

That’s not an impediment to use. Many people, myself included, don’t let their lack of knowledge about internal combustion engines stop them driving. As AI develops, and is used by more services and in more settings, it’s better to test public attitudes to scenarios in which they might encounter AI, as we do in the question below.

 

Ah. In every scenario tested a clear majority would prefer that the decision was made by a human, rather than AI. That sentiment is weakest where the volume of data involved is likely to be strongest – e.g. bank loan approvals or smart motorway limits – and strongest where judgement of some kind is placed on another human: guilt in a trial, a job offer, a university place, A&E prioritisation.


What does this mean?

None of this will stand in the way of AI, or offers any kind of serious barrier to its wider use. Humans are pretty bad at reacting to new technology, and pretty good at integrating them, more or less seamlessly, into daily life.

Some data from the excellent Archive of Market and Social Research offers a window into this. In the October 1997 edition of MORI’s British Public Opinion it was reported that…

  • only 22% of the public felt comfortable using the internet
  • 40% agreed that Information Technology was a threat to employment
  • On specific uses, only 38% said they would feel comfortable using a computer to shop – 45% said they’d feel uncomfortable shopping online
  • The public were evenly split on their comfort/discomfort with using a computer to bank, to access information on company or Government services and were decisively uncomfortable with receiving medical advice via a computer.
  • 64% worried about the effects of computers on children’s eyesight.

Analysts looking at these figures might have concluded that the internet and information technology would struggle to find an audience, or overcome these discomforts. I can’t find equivalent data on dishwashers but I’m prepared to bet people were just as cautious, just as sceptical. The technology is developing all the time, but our views about it don’t change at all.

None of this is to say that there aren’t real, serious risks from the development and wider use of AI.  The seriousness with which lawmakers around the world – especially the EU – are developing new regulatory frameworks is testament to that. But a hostile reaction to new technology is completely typical.

 


 

All this - and much more exclusive data on how consumers feel about AI - is contained in our new subscriber report, The AI CitizenIf you're not a subscriber, you can find out more about the membership here, or have a look at one of our previous AI reports, here.