Social Implications of AI

Background

Social implications generally refer to anything that affects an individual, a community or wider society. The social implications of artificial intelligence (AI) is an immensely important field of study, because AI technology is expected to steadily continue to permeate other technologies and, inevitably, our society as a whole. Although AI stands to bring about great societal benefits, it will certainly have its downside, as summed by the words of Ray Kurzweil, a futurist and creator of optical recognition technology who once said ‘fire cooks our food and also can burn down your house’. Similarly, the problem of ‘AI washing’ threatens to overinflate expectations for the technology, undermining public trust and triggering a backlash (Waddell, 2019). The tech sector’s ‘fake-it-till-you-make-it’ attitude plays into this problem. If AI is to truly be a force for good, it must have the capacity to develop from experience with people and learn from success and failures in social settings. Without self- and social-awareness, professional codes of ethics and increased levels of social and moral responsibility, society will never truly enjoy the benefits of AI (Whitby, 2003). Given that humans tend to use any notion of non-human agency as a convenient curtain behind which to hide their own culpability (Whitby, 1988), this may well serve the parts of society who would rather block the positive application of AI to protect their own self-interests, personal agendas and, most of all, economic and political power. With this in mind, we should ask ourselves ‘what type of society do we want’? Will people simply become consumers served by intelligent systems that respond to our every whim? Are we headed for a society of abundance (Blumenthal, 2011), which, despite requiring a co-operative culture (Linde and Neuvonen, 2017), is instead fast approaching a tipping point between convenience and dependency? Will society ultimately become a fusion of human, machine and information systems (e.g. trans-humanism) that forges a new level of co-existence and interdependence between technology and humankind (Yoshiyuki et al., 2014)? Trying to prevent or opt out of advancements in AI is a misguided and futile strategy.

Ai For Social Good

To analyse potential applications for social good, the McKinsey Global Institute (Chui et al., 2018) conducted a study in which they compiled a library of circa 160 AI social-impact use cases (see Figure 1) that could potentially help hundreds of millions of people in both advanced and emerging markets (EM).

Figure 1.

AI spans many distinct social domains

The study looked at domains of social good where AI could be applied and the most pertinent types of AI capabilities, as well as the bottlenecks and risks that must be overcome and mitigated if AI is to scale up and realise its full potential for social impact. It was divided into five sections:

• 1.

  Mapping AI use cases to domains of social good

• 2.

  AI capabilities that can be used for social good

• 3.

  Overcoming bottlenecks, especially around data and talent

• 4.

  Risks to be managed

• 5.

  Scaling up the use of AI for social good

Eighteen AI capabilities were identified that could be used to benefit society. 14 fell into three major categories:

• •

  computer vision

• •

  natural-language processing

• •

  speech and audio processing