How to evaluate the social impact of your AI strategy
06 August 2020
By Luciano Oviedo and Sotirios Paroutis
The range of topics discussed at the big tech antitrust hearings on Capitol Hill in Washington in July, from the regulation and censorship of speech to the commercial exploitation of user data, highlighted the pervasive impact on our lives of many emerging digitally-enabled technologies.
Increasingly, it seems that these technologies – from AI to automation and blockchain to the Internet of Things – are double-edged, creating significant challenges for business and society, as well as providing unprecedented opportunities for growth and prosperity.
Indeed, beyond the reputational hazards of appearing in front of a US congressional sub-committee, there are many reasons why organisations might wish to integrate a social impact assessment for the emerging technologies they develop and use into their strategy development. And, in doing so, move beyond the traditional, but many would argue outdated, focus on the standard strategy fare of financial and market competition oriented metrics.
It might be to fulfil a corporate duty to make a positive impact on society, as with benefit corporations in the US, for example. Or that executive decision makers believe the effects of emerging technologies on society so profound that it is the 'right' thing to do, part of new social contract that gives organisations license to operate.
Another possible impetus is the commercial considerations associated with consumer demands for greater ethical and social responsibility from product and service providers. At its most basic, given the additional burdens imposed by the introduction of data protection regulations following widespread, digitally-enabled, commoditisation of data and pillaging of privacy, it might simply be for risk mitigation purposes.
It is not indifference that prevents action, either. Our research indicates that executives are concerned about the effects of emerging technologies on society. Instead, the issue for many organisations is a glaring gap in their strategy toolkit, limiting their ability to make social impact evaluation a proactive part of operations, and the creation and development of products and services.
As a result, and with so many companies heavily investing in AI technologies and adapting business models accordingly, we have created the Social Impact Strategy Analysis (SISA) strategy tool. It was developed based on our experience of more than 30 one or two-day workshops, involving more than 500 internal and external stakeholders, at a Fortune 50 multinational technology corporation headquartered in Silicon Valley, California.
The SISA methodology can be used to evaluate plausible systemic risks and benefits associated with AI technologies (or other emerging technologies) and related products and services, across a range of social factors such as environment, energy and 'society' (ethics, economics, gender, culture and language). It is a flexible tool that can be used by emerging technology innovators and users, as an individual or group process, to investigate social impact at a number of levels, from individual communities to society more broadly.
Importantly, the SISA approach is an open and participative process designed to incorporate a range of stakeholder perspectives, including those of affected communities and citizens, before any deployment of an innovative technology enabled product or service.
Once the technology associated challenge or opportunity has been established, the SISA strategy exercise follows a number of distinct steps. These can be loosely divided into three stages: the initial setup; understanding the value ecosystem; and the social impact factor analysis.
The four steps in the initial set-up – participation, prioritisation, use case and content – cover the selection of participants, deciding on and prioritising the social factors to be analysed, clarifying and describing the use case, and gathering factor and use case related information to be used during the second and third stage.
Our methodology is designed to accommodate the analysis of up to eight factors relevant to the use case. It is important to select social factors that allow for a discussion about the dynamic interaction and interplay between the technology and factors (rather than just the one-way impact of technology on the factors).
The second stage establishes the context for the proposed use case using three separate questions: 'Where are we?' maps out the existing platform ecosystem landscape for the specific technology; 'where might we go?' explores the likely evolutions of that landscape; while 'why should we go there?' identifies potential benefits and opportunities available in that future landscape.
How to use the Social Impact Strategy Analysis tool
For autonomous vehicles (AVs), for example, the platform landscape might include elements such as safety and security, autonomy, intelligent manufacturing and on-board sensors. In turn the onboard sensors layer would contain sub-categories, such as location, vision/camera, and LIDAR and RADAR navigation systems.
Exploring the evolution of the landscape using two dimensions – the nature of the passenger-vehicle contract and type of driving control (for exmaple, human or autonomous system) – might suggest that a shift from ownership and human control to ride-sharing and autonomous systems is highly likely. Lastly, in terms of the potential opportunity, further analysis could outline the predicted economic impact of AVs (some $7 trillion in relation to consumer and business products and services).
Having established the platform landscape and the organisation's position within it, present and future, the next stage is to analyse the interaction between technology and selected social factors. For AVs, that might include energy and environment, with risk-benefit analysis considering the impact on emissions, the energy consumption of different AV technologies, and the energy and environmental effects of battery manufacture and disposal.
Analysis of an economics factor, might weigh potential benefits arising from reduced healthcare costs due to lower accident frequency alongside issues such as predicted job losses.
The final step is to summarise the aggregate results of the SISA exercise. A simple way of doing this would be using a risks/benefits scorecard covering all of the factors.
Adopting an open, design thinking approach to strategy development is a fundamental part of the SISA methodology. Rather than having small group, closed, top-down decision-making, the aim is to design an open middle-out, bottom-up process.
A range of stakeholders from the technology's ecosystem should be included, from frontline and mid-level staff outside the relevant business unit, but still inside the organisation, to suppliers, customers, regulators and government agencies.
One of the main benefits to this approach, besides accessing specialist knowledge and insight, is the potential for creating buy-in for a proposed strategy while, at the same time, identifying blindspots, pitfalls and maximising risk mitigation.
It is unlikely that any single tool will deliver solutions for all of the strategy challenges an organisation faces. However, adopting an open, participative, design thinking perspective alongside the SISA tool, should at least improve the overall quality of strategy development.
It provides organisations with a means of evaluating, representing and sharing the potential impact of interactions between technology enabled use cases and relevant social factors. All stakeholders in a platform ecosystem with access to this information can use the analysis to game a range of scenarios and inform relevant decision-making processes.
In a broader context, there is an ongoing debate about the evolution of capitalism, prompted by criticism concerning the fairness of capitalism in its traditional and dominant form.
If the status quo persists, the effects of introducing emerging technologies into almost every aspect of our lives may only serve to accentuate and exacerbate any problems with the current system.
Conversely, providing the people affected by these technologies with an opportunity to influence their development and use through open strategy tools such as SISA is likely to benefit creators, consumers and society more generally.
Knight, E., Daymond, J. and Paroutis, S. (2020) "Design-led strategy : how to bring design thinking into the art of strategic management", California Management Review.
Oviedo, L.C., Paroutis, S. and Smith, M.A. (2020) "So, you think you have an AI strategy? Think again", Industrial Internet Consortium paper.
Luciano Oviedo is Founder and CEO of tech start-up Tempugo WAITX after 12 years in new product development and corporate strategy at Intel. He is studying a Doctor of Business Administration in computational strategy.
Sotirios Paroutis is Professor of Strategic Management and teaches Strategy & Practice on the Distance Learning MBA, Executive MBA and MSc Marketing & Strategy. He also lectures on Strategic Advantage on the Executive MBA (London).
For more articles on the Future of Work sign up to Core Insights here.