What is AI actually currently capable of in the workplace?

Guest post by Roderick Crawford, Public Sector Director at SAS UK & Ireland on real use cases for AI, powerful outcomes, and how to achieve them.

The question of whether AI is friend or foe never seems to lose momentum. How will it impact jobs, life, the economy? While we debate the possibilities, however, AI is proving its viability. Governments are actively implementing digital transformation plans that include the use of advanced analytics and AI to drive improved services for organisations and individuals alike.

These ‘smart departments’ can be more collaborative, working in ways that are based on information sharing and analytical insights for a more intelligent, informed workflow and service. Add the fact that all this can be achieved for a fraction of the current cost and far more efficiently, and you can see why AI is becoming the dominant trend in many industries.

Rightly so, for we can already see just how disruptive a force it has been in the commercial sector: in retail (with Alibaba and Amazon), in transport (with Uber), in entertainment (with Netflix) and thousands more applications. So just how did these organisations get past the AI hype and put it at the heart of their very being? More importantly, how can government organisations leverage the advanced analytics that powers AI, infusing it with intelligence so that it radically and quickly changes service delivery and citizen outcomes?

What challenges is the public sector facing?

It boils down to financial and human resources as well as time, coupled with the growing number of citizens and organisations requiring interaction, many of whom have an increasing number of complex requirements.

By automating the analysis of the large volumes of data that are available to governments and learning from them, artificial intelligence can help with much of the day-to-day decision-making that is required. The generated outputs from this process form a feedback loop that can result in continuously improving services and more efficiency.

AI can also learn more about human behaviour, even at the individual level, than any single civil servant could possibly hope to. All of which means that governments around the world are beginning to adopt AI to benefit citizens in all manner of ways.

It’s important that government departments apply a structured approach and thinking with regard to AI adoption. Laissez-faire approaches might work for cash-rich development houses and venture capital-backed start-ups, but they are highly risky for public sector bodies with accountabilities to Parliament and to the public.

AI on the ground

Many government activities are ripe for AI adoption. Firstly, it’s interesting how defence departments use AI to analyse a far richer and broader suite of contextual data to detect and evaluate emerging threats. Law enforcement authorities use AI to monitor police officers, monitoring patterns of stress in order to balance their working lives and mitigate the risk of poor frontline decision-making. They also use AI to analyse connections between criminal entities by rapidly analysing the patterns of their activity and networks of associates. All of which helps frontline officers to get a handle on the threat landscape and deliver better outcomes for all.

Ministries for agriculture use AI to help farmers understand the best times to sow and harvest crops, using decades of weather data and current rainfall information – and there is no need for in-field sensor equipment and all its associated installation and management costs.

In education, exam marking can largely be deployed using AI, learning and improving over time and only requiring human input on creative subjects that require the kind of “thinking” AI cannot achieve. Critically, this frees up teaching staff for lesson planning and individual tutoring.

Many more decisions can be made faster on critical human outcomes such as welfare qualification and payments, immigration, environmental and urban development modelling to mitigate the effects of housing and infrastructure, given our changing global climate.

The FATE approach

Despite these many potential benefits, however, there are still widespread concerns over the impact of AI on the employees and communities it serves. In order to mitigate concerns that AI will dehumanise the provision of public services, organisations should adopt a code of AI ethics. These should be governed by the ‘FATE’ approach, under which AI should be fair, accountable, transparent and explainable:

Fair: It is essential that AI be devoid of bias and corporate discrimination. Human bias should be removed by the involvement of AI, not replaced with new machine-learned bias. It’s essential to question who decides on the accuracy of AI decision-making and ensure a diverse range of inputs at the design stage to avoid inbuilt discrimination.

Accountable: The ownership of AI-powered decision-making should be clear and accountable. AI should not simply be left on its own to make decisions on behalf of the organisation but partnered with a human project owner to ensure the resilience of the decision-making.

Transparent: The method by which insights and decisions are reached must be clearly laid out from start to finish in order to avoid ‘black box’ scenarios. Under GDPR, consumers have the right not to be subjected to decisions made by ‘unseen’ analytical processes. Greater transparency allows for greater trust in decisions made by the system.

Explainable: AI-generated decisions must be subject to explanation and make logical sense. A transparent AI process is more likely to produce models that can be explained. Organisations need to find the balance between accuracy and interpretability.

Putting our trust in the machine

Ethics are fundamental to the successful implementation of AI projects in the public sector. AI is not a threat: it can be used for good, to support communities, aid ecological and medical projects and improve their efficiency and accuracy. This in turn will allow civil servants to be released from laborious activities and redeployed to those that require lateral thinking, empathy and creativity, which are uniquely human traits.

Like all technological developments, AI will change the way we work. But just like word processors and mobile phones, they will evolve our workloads, not remove them all together. AI is not the job-stealer we fear. Workloads will change, not disappear. Public sector employees need to work alongside AI, both to ensure systems are meeting the FATE parameters and to decide on actions resulting from AI insight. AI will allow humans to be more human, and put our skills in communication and collaboration to best use.

Roderick Crawford is Public Sector Director at analytics and AI leader SAS. In his role he helps public sector organisations implement analytics to drive valuable insights and improve services.

If you would like to have your company featured in the Irish Tech News Business Showcase, get in contact with us at [email protected] or on Twitter: @SimonCocking