I think AI has huge potential in increasing productivity and making white collar workers more effective in their jobs. The main driver for this increase in productivity, in my view, is that AI will take on the mundane, repetitive tasks in our jobs, giving us more time to focus on the creative and difficult parts (see here for a discussion of some empirical evidence along these lines). But that also means that the boost to our productivity will vary from person to person.
In an experiment with telemarketers, the sales agents were measured how well they could answer questions from customers that were not part of the standard reference book. Was there a difference in how well telemarketers could answer difficult, out-of-the-box questions when they were assisted with AI?
In general, telemarketers were better at answering difficult customer questions when they were assisted in generating sales leads by AI than without it. This seems to be a result of the salespeople having more time to think about these questions when their bandwidth was not occupied with the admin work and because AI-generated sales leads were more comprehensive than the sales leads generated by the salespeople themselves. Hence, agents were better prepared for their calls with customers when the leads were generated by AI.
But these gains in effectiveness did not come evenly to all telemarketers. The researchers split the sample between top salespeople who managed to be in the top third of their company in terms of sales volumes and bottom salespeople (bottom third in sales volume) in their day-to-day jobs. Without the assistance of AI, the top salespeople were 2.5 times as successful in answering difficult customer questions than the bottom salespeople (and hence driving a higher sales volume because these difficult questions come predominantly from customers who are already seriously looking into buying the product). With the help of AI, the difference in success rates between top and bottom salespeople rose to 2.81.
Furthermore, top salespeople reported a more positive attitude towards AI and generally more positive emotions at work which in turn enhanced creative problem solution. Meanwhile, bottom salespeople reported more negative emotions at work when assisted by AI and made more limited improvements to their work processes.
And this last point is important, in my view. AI is a powerful tool that can make us more productive and effective at our jobs. But in order to do so, two things have to be present. First, AI helps more creative people more. People who are already good at solving problems and being creative benefit more from having time freed up from mundane admin work. Second, AI is not just any odd tool. It creates an emotional impact on its users. Some users can feel threatened by the power AI has, while other users may feel AI is a real blessing.
I think AI is what you make of it. Users need to approach the technology with an open, constructive mindset. People who are sceptical of technology and feel threatened by it will not use AI or use it less effectively. The result is that eventually, this emotional obstacle will contribute to them lagging more and more behind their peers who embrace the technology.
from BNP PARIBAS' publication:
Global economic impact of AI: The future is coming faster
-Adoption can be relatively fast. Generative AI is likely to transform economies faster than previous episodes of significant technological change. We could see mass adoption among corporates in the coming two to five years. We estimate a 1% productivity gain already by the end of the decade.
-Readiness varies across countries. Advanced countries look better positioned than emerging countries to harness the benefits of AI, according to our country ranking of AI readiness. (Italy is the worst equipped)
-Impact is likely disinflationary. The disinflationary impact of AI could be around 1% a year, especially on services prices, helping to offset other potential long-term inflationary forces.
-While generative AI is likely to boost productivity, its effect on the workforce is less clear-cut. Over the long term, history shows that productivity gains create their own demand for labour, which will support employment. In the short to medium term, large job disruptions look likely, especially in the service sector. Recessions usually trigger corporate restructuring and labour adjustments