The advent of artificial intelligence (AI) has ushered in a new era of possibilities across industries and job functions. While AI can unlock untapped potential for both organisations and individuals, it comes with accompanying risks, like algorithmic bias and privacy breaches. With countries such as Singapore shifting their focus towards bolstering responsible AI adoption, there has been increasing demand for more stringent AI development, application and regulation. However, committing to AI development and adoption is a significant financial and operational investment. Organisations are now faced with the dilemma of deciding if committing to responsible AI will be a cost or strategic investment.
The fundamentals of responsible AI
When ensuring responsible AI development and deployment, it is important to consider principles of transparency, explainability, trust and accountability. It is also important to ensure human agency and oversight to promote fairness and inclusivity.
We explore how Singapore can further support its startup ecosystem as an AI hub
The foundational elements of responsible AI are explainability and transparency. For ease of use and accessibility AI processes and decisions should be made clearly explainable. This helps users better understand how AI algorithms make decisions which eventually helps cultivate trust towards AI solutions.
Trust and accountability go beyond mere compliance. They form the fundamentals of how users interact, adopt and eventually trust AI solutions. In line with the Association of Southeast Asian Nations’ (ASEAN) guide on AI governance and ethics, developers must ensure data is handled sensitively and securely so users trust AI models with private, sensitive information. AI solutions must also adhere to the highest standards of privacy, security and ethics.
While AI greatly increases efficiency, the human touch is irreplaceable. AI solutions should supplement human judgement rather than replace it. In line with the creation of national AI councils to guide the ethical adoption of this technology in the region, AI solutions based on human values and ethics should prioritize human oversight, giving humans the last say in decision-making. Just as how a car is a machine, we use to shorten our travelling time, AI serves as a tool to support us in our endeavours to progress as a society.
Today where there is a continual pursuit for progression, diversity, fairness and inclusivity are important values that we cannot ignore. When developing responsible AI, we need to ensure these values and corresponding data points are fed into AI development, to ensure the underrepresented in our communities are accounted for, valued and represented.
The cost argument
Developing and implementing responsible AI undoubtedly requires resources. This includes both direct and indirect costs to organisations. For direct costs, organisations need to hire the right personnel, like AI specialists, and existing staff need to be trained in how to develop the skills necessary for effective AI implementation and management. Organisations also need to ensure the current IT infrastructure is equipped and advanced enough to handle high computational loads and data storage requirements of large language models (LLMs) that AI models utilise. Should organisations choose to engage a vendor to help with AI implementation, there will be additional costs in areas such as model development, testing, auditing and monitoring of AI models. These help to keep AI models updated, and compliant with regulatory requirements and ensure potential risks are identified and mitigated in a timely manner.
Indirect costs to responsible AI development include the opportunity cost of time and resources spent on AI instead of other projects. With the constant development and improvement of AI, regulations and best practices are constantly evolving with the introduction of new frameworks within the region like Singapore’s AI Verify toolkit. This can lead to challenges in staying abreast with the latest AI development and additional costs to organisations. Should AI solutions result in negative press or legal issues, organisations also need to be prepared to provide explanations and allocate resources to rebuild trust and compensate for losses. Beyond that, expenditure on communication and transparency will also be necessary to gain and keep the trust of customers.
The investment argument
We’ve already seen how much untapped potential AI has unlocked. AI tools can automate tasks, optimise processes, and generate insights. Responsible AI as a strategic investment can unlock a multitude of benefits which include enhanced brand reputation, improved internal work efficiencies, and increased employee morale which ultimately drives long-term business sustainability.
By informing end users about the integration of AI into business processes and workflows, the organisation presents itself as being ‘tech-forward’. Responsible AI practices promote confidence and openness among stakeholders by serving as guardrails to AI development and application. This in turn, enhances brand loyalty and recognition.
AI may also assist internal departments in keeping abreast of recent regulatory changes, proactively identifying irregularities, and generating clear visuals that enable more accurate and reliable insights. This enhances overall improved decision-making, boosts internal efficiencies and gives organisations a competitive advantage in the market.
Equipping employees with AI knowledge and competencies fosters a culture of continuous learning. This culture empowers individual employees but also demonstrates the organisation’s commitment towards continual development. Having a human-in-the-loop approach in an organisation’s AI strategy also ensures accountability for AI-generated outputs. This ensures supervised but efficient workflows which creates a ripple effect, where business units can individually harness the benefits of AI, fuel innovation and contribute towards long-term business sustainability.
Finding the sweet spot
While organisations recognise the potential benefits of responsible AI, when investing in it, it is key to quantify its impact. Some key performance indicators (KPIs) that organisations can consider when measuring the outcome of responsible AI include customer satisfaction, employee turnover, and financial performance. Business leaders can also conduct a cost-benefit analysis to evaluate the financial effects of investing in responsible AI practices.
Ultimately, it is key for organisations to find a sweet spot on how best to leverage AI which is unique to its business offering. This must be done concurrently whilst balancing between the various pros and cons of adopting responsible AI. With the increasing focus on AI as a pivotal player in enabling progress in Southeast Asia, CIOs and business executives in the region must establish clear metrics, indicators, benchmarks and standards to track and ensure the efficacy of AI technologies. Adopting and utilising responsible AI is an ongoing process that requires continuous monitoring, assessments and adjustments. Starting AI initiatives as pilot projects with experienced industry partners or academic institutions can also be another way organisations pick up industry best practices and concurrently reduce implementation or adoption costs.
While responsible AI may require a significant upfront investment, the long-term benefits include enhanced brand reputation, improved internal work efficiencies, and increased employee morale which drives long-term business sustainability, making it a strategic tool that organisations should consider adopting. With a well-developed, measurable and responsible AI strategy, organisations can tap into AI to drive impactful and sustainable business outcomes while minimising its costs and risks.
The article titled “Responsible AI: A cost or a strategic investment?” was authored by Philip Madgwick, Regional Vice President of Alteryx Asia & ANZ
About the author
Philip Madgwick is the Regional Director – ASEAN for Alteryx, based in Singapore. Philip and his team work closely with their clients across the region, enabling them to turn their data into actionable insights as part of day-to-day activity or a wider digital transformation.