IBM: The principles of effective AI

Artificial intelligence (AI) can spark fears for leaders. Not only, is it easy to jump to a catalog of sci-fi films that showcase machines rising up and enslaving humanity, but AI has also been accused of creating biased hiring processes and career paths.

Despite these concerns, AI is a developing technology that can genuinely help businesses. To understand exactly how AI and other technology can empower the workforce, UNLEASH sat down with Anshul Sheopuri, vice-president and chief technology officer at IBM Workforce.

Before we jump into the world of data, Sheopuri notes what his role entails: “The way I like to think about my role is I get to envision the future of the IBM workforce”. And then he has “the exciting job of making it happen”.

On a more formal note, he states that his role is “having an integrated data platform that connects across the employee journey and ties to dice the business, AI as a silver thread across an employee journey, ultimately resulting in personalized experiences very similar to what you would have in consumer-grade experiences.”

Make employee experience a positive journey

Let’s jump into what AI actually looks like inside tech giant IBM. 

Sheopuri says that he looks at “consumer-ability” rather than “usability” when evaluating workforce applications at IBM.

He explains: “To me, the notion of consumption is much more upfront, a user problem requires understanding what the users really want versus what they might say that they want.” 

Anshul Sheopuri, vice-president and chief technology officer, IBM Workforce

Unpacking this further, Sheopuri says that it is essential to understand what users want even if they are unclear about it themselves.

Once the want has been realized, leaders can create an intersection between the business and employee needs.

No matter what employee issue your organization is trying to overcome, Sheopuri recommends that you achieve clarity of the business problem. On top of that, he notes that it is essential to research and begin testing what works. 

Of course, not every tool or idea will be well-received, with that in mind failing fast is essential.

On top of that, Sheopuri claims it is important to have a “continuous loop of iteration and feedback, in order to make sure that you’re driving towards the right consumption that scales with the impact of the experience and then results in the talent outcomes”.

Feedback and continuous assessment are key to Sheopuri and IBM because of the diversity of its employees.  He explains: “It’s very easy to design for the majority – if you look at only survey results, and treat them as quantitative data – but when you hear those comments, you then get really diverse perspectives.”  

Leveraging AI

AI is an essential component in Sheopuri’s aforementioned “consumer-ability”. He offers an example of how learning and development systems work within IBM.

To help employees develop skills and forge career paths, IBM leverages a digital platform that helps identify personal learning paths based on existing titles, skills, and interests.

Sheopuri likens the process to Netflix; “Think of it just like that, it gives you different channels of this popular learning that’s out there, there are other people who are in a role similar to you, and this is the learning that they are doing”.

This intelligent learning system also helps IBM employees find new roles at the company based on their career paths and skill sets. This not only helps IBM internally but the employee can evaluate their progress and decide how they want to further their career externally or internally.

Sheopuri concludes: “That’s an example of how we’ve put AI into practice with results that solve a user problem like how do I improve my career as well as a business problem because that results in reduced voluntary attrition which is great for the company”. 

To ensure that AI works the way it should, Sheopuri shares: “We have as four pillars of trustworthy AI. And that’s sort of a common recipe and template we take across all AI that we developed”.

Transparency

The first pillar is transparency so that employees can see their progress and make their own choices with the help of AI. 

To make sure AI stays on the right track, IBM looks at its “nutritional value”. By this Sheopuri means looking at what sources are recommending avenues of growth for employees, and how impactful they are.

Explainability

Transparency ties into what Sheopuri describes as the second pillar; explainability, which is about showing staff why they should use AI career paths.

For example, the AI will let employees know what other learners in similar positions are training in, and the courses that are becoming popular. This allows employees to be confident about the training they choose as they carve out their future roles.

Fairness

The third principle of AI for Sheopuri is fairness. Making sure AI is unbiased is essential, after all, systems mistreating employees can lead to widespread issues and even court cases.

As a result, IBM looks at data queries like “what is the percentage of men versus women, different traverse groups? who’s consuming learning and what content and what outcome is that producing?”

On this note, Sheopuri discusses how diversity and inclusion goals can be enabled through AI: “It’s an opportunity to give equal access to options of opportunity to different groups because AI enables scale.”

Privacy

Sheopuri explains that the final element of AI is privacy. This means making sure that the right people have the data, as well as guaranteeing people’s information remains safe.

Despite these elements being essential, Sheopuri adds: “The only other thing I would leave you with is this is not about a checklist at the back end of deploying an AI. It’s something that you embed upfront in the design process as you co-create with users.”

Establishing common standards

Although there are huge benefits to using AI at work, Sheopuri notes: “What we have to acknowledge, as practitioners in this in this space is, we have not seen the emergence of common standards.”

Everyone has their own definition of AI and IBM has been on a mission to work with companies across industries “to create a precursor to standards, a common set of questions we should be asking in the workforce space”.

The idea is that there will be an “emergence of a common vocabulary across different companies. Because once you have that common vocabulary, with working with different vendors with different companies, we can begin to see that nutritional value, much more transparency.”

Once standards are set, there is a foundation “for AI being deployed at scale in a manner that gives equal access to opportunity”.

Finally, Sheopuri considers how AI could be history’s latest innovation. He places AI’s potential alongside technological inflection points such as automation or the establishment of cloud that caused computing costs to drop.

Rather than have fragmented data across disciplines in an organization, Sheopuri believes that AI can unite data and give insights into employees while solving issues at speed and scale. 

Undoubtedly, this would change the world of work as we know it.

Want to hear more from UNLEASH? Sign up now for free access to all our online content, exclusive reports, as well as discounts for our events.