Why it’s crucial to build confidence in AI

Transparency is something that people expect of organizations today, regardless of AI. From practices and products to how businesses engage with employees and customers, transparency is how an organization becomes a reputable partner, employer and goods/services provider.

But organizations can’t always be completely transparent. They can’t legally share confidential data, for instance, and wouldn’t want to divulge information that could damage their business. So firms already have to balance how much to share with their customers and stakeholders to maintain relations without giving away what makes the business competitive. According to Kumardev Chatterjee, CEO, technology innovator and entrepreneur and WEF recognized expert , it’s the same in the age of AI.

“Just because a company is AI-driven, it doesn’t mean that what is illegal to do today will become legal tomorrow,” Chatterjee explains. “Organizations aren’t going to start putting their employees’ confidential information online just because of AI. And just because it’s AI-driven doesn’t mean that an organization that has the right balance has to be more transparent than it is today.”

In Chatterjee’s opinion, the same issues will persist – just in an AI context. “As usual, it’s a question of communication, audience management and customer experience.” The only difference, he says, is that “AI will simply provide organizations with better tools to perform these tasks – such as helping organizations to better understand customer sentiment and so develop better customer experience strategies.”

The next question is how leaders can create AI processes that are ethical, safe and unbiased. For Chatterjee, the solution is twofold.

First, companies will need to hire for a new role – an AI auditor. “This will be an independently trained, legally and technically aware individual,” Chatterjee explains. “Someone who can tell organizations whether the processes they’re following are safe.”

Second, companies should address the larger question of AI practices within their industry. For this, Chatterjee believes it will be essential to have industry-led AI transparency standards. 

Organizations have to be able to explain how they get results from AI processes. This helps their employees and customers to see AI as a useful tool instead of a malevolent, destructive force. But is it possible to fully explain AI processes?

“The short answer is no,” says Chatterjee. In his experience, when customers ask companies to tell them how they are doing X and Y using AI, what they really want is reassurance – they want to know how the organization is making sure that their lives will not be compromised by its actions, processes and software.

This need for reassurance has surfaced every time a new technology has entered the market, and it will keep coming up. “Today it’s AI, tomorrow it’ll be quantum computing,” says Chatterjee. "The next day, it’ll be laser-based recognition.”

A good example here is airport scanners. When they were introduced, people were appalled that these machines would reveal their bodies to airport security. “Now, you don’t even think about it,” says Chatterjee. “In fact, people get upset if an airport doesn’t have a body scanner, because it takes more time to go through security. The same will happen with AI.”

So instead of explaining all their processes in intricate detail, organizations will need to balance what they do and don’t explain to their customers – just like they do with transparency.

Leaders must provide reassurance that personal data will not and is not being compromised, that there’s no inherent bias, and that all digital processes are operating ethically. But the organization will also need to remind clients and customers that it can’t always disclose how it reaches a certain outcome if it would be competitively disadvantageous. “This is a tool,” says Chatterjee. “And like any other tool it’s our responsibility to make sure it’s used responsibly without compromising the customer or our business.” 

“AI turbocharges collaboration,” says Chatterjee. This is because a lot of the power of AI comes from its ability to process and analyze vast bodies of data. But in many cases, the data an AI system needs lies outside the boundaries of a single organization and is held by partners, suppliers, customers and others in the organization’s wider ecosystem.

Is the cloud the answer? “Most of our data is on the cloud,” says Chatterjee. “And AI doesn’t discriminate between which network data sits in. It only needs the data.” Going back to transparency and explainability, this doesn’t mean the AI has access to every bit of data. “Although AI can go across organizational borders,” says Chatterjee, “it’s up to leaders to decide what they’re willing to make publicly available.”

Let’s say an organization is planning to create a consortium to reduce its industry’s carbon footprint. AI tools would enable it to find, for instance, 20 organizations that have listed their carbon footprint and other relevant information on the public cloud. Once the consortium is formed, these organizations can then use AI to identify issues, set an industry standard and manage the carbon footprints of their supply chains. 

So with the rights tools and approach, data can help organizations to collaborate effectively. According to Martin Wezowski, Chief Futurist and Head of the Future Hub at SAP, sustainability is one area that’s already benefitting from this approach through SAP and Fujitsu’s societal resource planning (SRP).

Like enterprise resource planning (ERP), SRP uses AI, machine learning and cloud technologies to connect internal operations with external partners. But while ERP is more focused on operations, SRP “enables businesses to follow their purpose while driving business growth,” according to Wezowski.

“Let’s say a company wants to meet its sustainable development goals for clean water (SDG 6),” he says. “How could it know about, say, three guys in Kyoto who have developed an innovative filter with nanotechnology?” In Wezowski’s experience, looking for experts like these can cost businesses a great deal of time and money. And costs can continue to accumulate after a collaborator has been found because of different regulatory demands or company values.

Wezowski says that SRP could be the solution. It doesn’t just connect leaders with difficult-to-reach innovators, but through tools such as AI and Generative AI it also identifies potential risks early. “It’s a new era of enabling people to connect to a larger society of innovators,” says Wezowski. “Augmenting them through an autonomous network so they’re the best they can be.”