The future of work: AI, creativity and skills

So many conversations about AI and the future of work start and end with the concern that robots will take our jobs. Recently, what passes for informed comment is often little more than the new cliché: AI will not take your job, but someone using AI will.

However, the situation today, never mind in the future, is already way more nuanced. If you picture AI as just a job-stealing monster, try a different perspective: think of it more like having an incredibly efficient assistant, taking care of all those tedious tasks so you can use your brain for the interesting stuff: unleashing your creativity and problem-solving skills.

Xueming Luo of Fox Business School ran a real-world experiment in a high-pressure environment: a telemarketing company. You can picture those giant call centers, people racing the clock to finish calls quickly, following scripts, the works. Call centers are not exactly known for creativity.

So, they brought in AI chat bots to handle initial customer contacts: sales calls about a new credit card offer.

The task involved both AI and human agents reaching out to customers. The AI was trained to handle routine interactions, such as gauging customer interest in a promotion or offer. The primary goal of this stage was to confirm customer interest, which is typically a straightforward and repetitive task.

The study used a large sample size of over 3,100 customers, who were randomly assigned to different groups to ensure that the results were reliable.

The customers were divided into groups where they were either approached by AI or by human agents and the human agents were further categorized into two types: top-performing agents (experienced) and bottom-performing agents (less experienced or new hires).

This setup created four combinations:

• AI handling the task

• Top agents without AI assistance

• Bottom agents without AI assistance

• Human agents in the second, more complex stage

Measuring creativity

I said earlier that call centers are not known for creativity. Interestingly, it is precisely this property that Professor Lo and his team were interested in.

The experiment aimed to measure how creative the agents were in handling customer queries, especially when dealing with questions outside the standard knowledge bank. They used Natural Language Processing to analyze the conversations between agents and customers. If a customer’s question was outside the predefined knowledge bank (e.g., asking if a credit card could be used for a down payment on a house), it was considered a more complex, creative task.

Creativity was measured by determining how often agents successfully resolved these untrained, novel questions. The ratio of these successful answers to total questions was used as the "creativity ratio."

What happens to the humans?

Because the AI handled those more straightforward calls, it freed up the human agents to focus on the complex stuff: answering tougher questions, coming up with solutions for unique customer problems, and so on. So, instead of getting bogged down in those repetitive tasks, they had more headspace to be creative.

The results?

The experiment found that agents who worked with AI were 2.33 times more creative in solving complex customer questions.

The increase in creativity was even more significant for top-performing agents. These experienced agents were able to leverage the AI to handle routine tasks, giving them more time and energy to focus on creatively solving complex problems.

While both top and bottom agents benefited to some extent from AI assistance, the top agents showed a more pronounced improvement in creativity.

But the AI didn't magically turn every single employee into a creative genius overnight. And that's even more fascinating. This effect was amplified for the experienced top-performing people. They saw their creativity take off: it supercharged the people who were already good. However, some of the less skilled workers felt threatened by the AI.

AI had a significant, differential impact based on prior skill level.

This leads to an important question. How do we ensure people have the skills to thrive alongside all these advancements?

Supporting lower-performing staff

There's a real challenge here. Perhaps for you, dear reader of this highly intellectual newsletter, AI will be a gift to your creativity. But I think for lower-performing staff we can adopt the following strategies:

• Upskilling and Training Programs

  • We can offer specialized training to improve skills and confidence in AI tools, helping users understand how AI can be a valuable assistant rather than a competitor. As part of this, we can incorporate training focusing on developing creativity and complex problem-solving skills.

• Mentorship and Knowledge Sharing

  • It's often very helpful to pair lower-performing staff, or just new hires, with top-performing agents in mentorship programs. This allows them to learn strategies and techniques for handling advanced business scenarios. We could encourage regular knowledge-sharing sessions where top-performing agents share their successful practices and creative solutions.

• Phased Introduction to AI Tools

  • I see a lot of promise in the phased introduction of AI tools to lower-performing staff. Start with simpler applications and progressively move to more complex integrations. This can help them build confidence and familiarity with the technology, but hands-on practice sessions, where agents can interact with AI tools in a low-pressure environment, also demystifies the technology and shows its potential as an aid rather than a replacement.

• HR Support

  • I don't think we can overlook that AI does present a real fear to many workers. Hold open discussions about AI's role in the workplace to address fears of job replacement. Include staff in the conversation about AI deployment, especially as you adopt a phased introduction. Make it clear that AI is there to support and augment human capabilities, not to replace them.

• Redesign the Roles

  • We shouldn't think of AI as replacing a job so much as refactoring the role. We can modify roles to ensure that AI is seen as a tool for handling repetitive tasks, giving staff the chance to focus on customer interactions that require human touch and creativity. To do this, clearly segment tasks between AI and human agents so lower-performing staff understand which tasks they are responsible for and how AI can assist them in those tasks.

Professor Luo's research offers a glimpse into a more nuanced reality of AI in the workplace rather than simply replacing jobs.

When used effectively, AI enhances creativity and problem-solving, particularly among top performers. However, this differential impact also serves as a reminder that not all employees will experience the benefits of AI equally.

You can read Professor Luo's work here: https://www.fox.temple.edu/directory/xueming-luo-tuf35198