Council Post: Responsible AI for Responsible Leadership

Bill Vaughan once said that, “The price of power is responsibility for the public good”. In today’s world, artificial intelligence can be termed as a currency—one aspect of which drives innovation and productivity while the other aspect evinces concerns surrounding biases, privacy, replacement of humans by automation and more. The ever growing momentum of these concerns is further proof that the existing guidelines do not address them adequately. However, that gap may be bridged with ‘Responsible AI’ that aims to address these concerns and create solid frameworks on accountability, aiming for interpretability, fairness, safety and privacy.

Companies today are adopting a more extensive approach where responsible AI is the foundation of corporate social responsibility (CSR). They are trying to make the transition that responsible AI today is not only a technological problem but also an organisational one. According to the companies surveyed for MIT SMR and BCG’S report (2019) on AI—an estimated 90% have made technological investments owing to a significant rise in AI adoption across various industries and sectors. Organisations are now implementing AI to enhance businesses internally so as to improve external client relations by making their search experiences personalised, predicting demand for products, developing internal efficiencies and much more. Though the adoption of AI has witnessed an upward trend, the same cannot be said for organisations. The report also suggests that though 52% of organisations have an Responsible AI  programme, 79% report that the programme is limited in scale or scope—meaning, RAI initiatives are lagging behind strategic AI priorities.

Here are a few potential reasons for the limited implementation of RAI in organisations:

• Complex interpretation

RAI is a relatively nascent field due to which several organisations have only used the term inconsistently so far. There are interchangeable terms like ethical, trustworthy AI that are used instead and “Responsible AI” often becomes an umbrella term of which these other terms become key components.

•  Organisational challenges

The technical complexities are less as compared to internal organisational hurdles. The most common factors preventing the scaling of RAI are expertise, proper training and awareness among employees, and funding and centralised leadership. There is an urgent need to bridge this gap between aspirations and reality by adopting a strategic approach towards RAI. Tactical initiatives undertaken by RAI leaders could address the potential risks of such technology by evaluating the impact it has on its users.

• Empowering RAI leadership

There is a need for a central authority to be appointed to issue RAI initiatives. The leader would bring together stakeholders, identify talents across organisations and bring in policies that would help in creating AI systems. However, only a centralised authority may not be the answer, rather an organisational ownership where diverse perspectives are more welcome. 

• Connecting RAI efforts to CSR

There are two different rationales for combining RAI and CSR:

• “Fundamental alignment between RAI and CSR initiatives” experts believe that  for an organisation, tying in its AI efforts should feel natural given that bias prevention, transparency, and fairness are already in line with the core values of corporate social responsibility.
• “End goal of linking RAI to CSR”, from an organisational standpoint, in order to direct responsible business outcomes, the practical implementation of the responsible AI programme needs to be closely integrated with corporate social responsibility efforts. In order for digital transformation efforts to be effective, they must be in support of business drivers, such as corporate social responsibility priorities, that will enable and safeguard a company's ability to operate in the future. 

On the contrary, there is also a fear among experts that linking them both can be a risk. These companies may show that creating and integrating ethical AI properly is not merely a bonus and that doing so could be a substantial liability to business operations by launching corporate social responsibility initiatives.

To consider the two rationales, the organisation must be aware of its ability of innovation, digital transformation and the internal culture. 

• Business benefits

The benefits include improved products and services, better product differentiation, rapid innovation, increased recruiting and retention, increased customer engagement, better profitability and well preparedness for emerging regulations. These initiatives will also help reduce any kind of risks with growing AI maturity and as AI applications are deployed the failures decrease. Prior to AI maturity, adopting RAI maturity will reduce the risks significantly—especially with scaling—and will help identify AI failures. Organisations with better RAI programmes are likely to be more prepared to tackle changing AI regulations.

Leadership recommendations for the future 

•  Give up on checklists: Having an RAI may make organisations feel prepared for future opportunities and risks. A matured RAI is not driven merely by regulations and prevention of risks. But, the RAI needs to be tapered according to the organisation’s visions, cultures and CSR activities.

• Expand your view: Assess all the internal and external stakeholders for the adoption of AI. There is also a need to evaluate the impact your business will have on the society as a whole. This is where connecting CSR efforts to RAI would work within the organisation.

• Quick start: Most common obstacles—including lack of training and expertise—can be solved if RAI efforts are launched at an early stage.  An early start can help in identifying the problems and solutions to be catered to in the beginning rather than later for the organisation.

• Investment: It is important to invest in RAI programmes from scratch starting from budget to talent branding which must be encouraged. There needs to be funding even for social aspects like education, awareness, training programmes and much more along with a diversity in the people who are taking up these roles—including executives at the highest level.

RAI is not simply about being responsible because of a special technology. RAI aims to connect to broader corporate ideals and in becoming responsible citizens. RAI leaders cannot limit themselves to only risks but also need to ensure the growth of their goals and principles—which would eventually also affect their stakeholders. Leaders need to frequently network with AI development teams to make sure that they create and responsibly use AI. As Oren Etzioni said “AI is a tool. The choice about how it gets deployed is ours.”

This article is written by a member of the AIM Leaders Council. AIM Leaders Council is an invitation-only forum of senior executives in the Data Science and Analytics industry. To check if you are eligible for a membership, please fill out the form here.