If you were to entrust artificial intelligence with one of these tasks, which would it be: arriving at a complex medical diagnosis based on a series of symptoms and underlying issues, or finding the perfect employee for a new role? You may be surprised to know that if you were to answer the latter, BanklessTimes.com can reveal that you would be in the minority. A study surveyed participants’ reactions to artificial intelligence across 17 different countries, and found that, overall, people trust AI significantly more for use in the healthcare industry (44%) than they do in the human resources field (34%). There could be a number of reasons for this trend, but perhaps the most obvious is that while people tend to place more faith in the capability and helpfulness of AI systems, there remains a healthy degree of scepticism regarding their ability to act fairly.
There is an element of truth to this, and a number of recent studies have shown how human bias is often translated onto algorithms. In 2015, Amazon came under fire when the AI program it built to review job applications discriminated against women for technical roles, while in 2022 the UK-based Information Commissioner's Office commenced a three year investigation into whether AI job application software in the UK showed racial bias. While we will have to wait a little longer for the outcome, it isn’t hard to see how headlines such as these can impact the public’s trust in the use of AI technology in the HR field.
AI will revolutionise the NHS
Conversely, the AI medical industry has one seen favourable headlines in recent years, with AI’s ability to make calculated diagnoses heralded as a triumph that will revolutionise the crumbling NHS. A £36 million boost will speed up diagnoses for lung cancer, as well as fund the build and operation of a new tool that allows doctors and nurses to diagnose heart attacks more accurately. Although both of these projects are currently in the planning stages, similar and revolutionary AI software has already benefited 17,000 stroke patients and 25,000 diabetes patients since its introduction in September 2020.
One of the key findings of the joint study is that people are more likely to trust AI where it is perceived as having a direct benefit to them, and that is of course particularly true of the healthcare industry. In other fields, AI is trusted by a significant number of people when used in security (40%), and in recommendations (36%). Interestingly, however, even those who would entrust recommendations to AI would have reservations about sharing personal information in order to get results.
People in emerging economies are more likely to trust AI
The joint study highlights a significant variance in attitudes to AI across different parts of the world. India was the country most likely to trust AI for use in the healthcare industry, and Finland the least. This is an unsurprising result, because while 67% of people across the world admit to feeling optimistic about the growth of AI, the issue of trust is more divisive, with India reposing the most trust in artificial intelligence as a whole (75%) and Finland the least (16%). By comparison, 40% of Americans trust AI, whereas the levels for England and Germany are 34% and 35% respectively.
India’s trust in AI is largely replicated by the other emerging economies included in the study (Brazil, China, India and South Africa), all of which are more likely to feel that the benefits of AI outweigh the risks. These countries are also the countries most likely to already be using AI in places of work, with 77% of Chinese workplaces currently using AI – significantly above the worldwide average of 33%.
The Chinese are the most optimistic about AI
It is perhaps this clear understanding of what AI can do and achieve that makes the Chinese the most optimistic about its implementation. While India is the country most ‘excited’ about AI, its participants were also the most ‘relaxed’ about its introduction yet also the most ‘outraged’. This wide spread of contrasting emotions is very different to France, where participants tended to be more downbeat on the whole, with the highest incidence of feeling ‘worried’ and ‘fearful’ about adopting AI.
France certainly isn’t alone in its pessimism about artificial intelligence, because, worldwide, 61% of people admitted that they are either entirely unwilling to trust AI, or just feel ambivalent about it all.
Finland is set to be a leader in the AI field
Finland’s claim to have the least trust in artificial intelligence does not always translate into practice, however, because AI use in Finland is 39% and therefore above the worldwide average of 33%. 95% of Finns also admitted to being aware of AI prior to taking the survey. In fact, the Nordic country has plans to become an AI powerhouse, and with brands like Nokia at the helm it is leading a quiet revolution. In May 2019, Reaktor, alongside the University of Helsinki, created a free online course, Elements of AI, available to all EU citizens. The aim was to make AI accessible to all, not just a few elite coders, and the initiative has certainly borne fruit: over 40% of those who enrolled on the course are female, which is double the computer science average of 20% for similar courses.
Jonathan Merry, CEO of Bankless Times was surprised that with such a high level of public education surrounding AI, the Finnish still report low levels of trust and only moderate acceptance towards it.
In this instance, it seems that being better informed about artificial intelligence doesn’t necessarily lead to greater trust and acceptance. Finland really pushed on the need for an independent regulator to oversee AI adoption, rather than a government body, so perhaps the key issue is not whether the algorithms are biased or whether the software is faulty, but whether we repose trust in those who deliver it, vet it and regulate it.Jonathan Merry, CEO of BanklessTimes.com
50% of people feel that national universities are most trusted to develop and use AI. We’ve certainly seen this in Finland, with the University of Helsinki educating all EU citizens about AI with the notion that greater understanding equates to less fear. In the health sector in particular, national universities have been a force for good, with Stanford University developing algorithms for detecting skin cancer and the University of Oxford developing AI algorithms to analyse MRI Scans to predict which patients are at risk of heart attacks and strokes.
49% of those surveyed also feel that security and defence forces could be trusted to develop and use AI. Again this is a field in which we have seen real developments, with the United States Department of Defense’s Project Mavern using AI to analyse drone footage, and the Israel Defense Force developing ‘Ofek’, an AI-powered predictive analytics system that uses data from social media and news sources to predict terrorist attacks. Again, the common thread is that the public are keen to put their trust into AI in fields where there is or might be a direct benefit to them.
In terms of who is least likely to be trusted to develop AI, commercial organisations and governments have that unfortunate distinction, at 28% and 33% respectively, and this is particularly problematic given the increasing use of AI by both government and business.
For AI to flourish in the future, we at Bankless Times think there needs to be close collaboration between governments, universities and businesses, overseen by an independent regulatory body that the public really have faith in. As attitudes around the world begin to soften and the public start to witness the technological leaps forward that can be achieved across all fields of life by a responsible and well-regulated AI industry, it is only a matter of time before trust in artificial intelligence becomes the norm rather than the exception.
Contributors
