A future where Artificial Intelligence (AI) renders humans largely redundant in the workplace is one view of AI’s societal impact. Another view is that although AI will redefine the concept of earning a living, the technology will liberate individuals from routine tasks and enable them to perform to much higher levels.
Both views were discussed at MIT’s AI and the Future of Work conference, on November 1-2, 2017.
Several speakers offered current examples of how AI is augmenting, not replacing, human talent.
Thomson Reuters has introduced AI technology that identifies breaking news stories on social media. But the tool has not eliminated journalists from the news reporting process. Rather, journalists use their domain expertise to analyze AI-generated stories, and spend more time on developing ideas for articles.
Veterinary products and services company IDEXX uses AI to help veterinarians analyze urine samples from animals. The system can analyze microscopic images and diagnose diseases as accurately as human specialists can. However, the system is a support tool for veterinarians – not a replacement.
Similarly, AI systems can now “read” x-ray images, but radiologists must sign off on diagnoses. Also, the technology allows radiologists to devote more time to patients.
AI is augmenting human expertise in the supply chain domain. For example, the technology’s ability to identify patterns in extremely large data sets is being used to spot data anomalies that could flag inefficiencies that need to be addressed. This super-human pattern recognition capability also is being harnessed to analyze customer buying behavior and refine demand forecasts.
As AI applications expand, the technology’s role as a support system will likely grow. Imagine, for example, advanced AI assistants in truck cabs that help drivers to meet tight delivery windows.
Still, as AI technology evolves and becomes more sophisticated, the chances are it will outgrow these support functions. Advances in deep learning – the application of neural networks that mimic the human brain to give machines high-level learning capabilities – will equip AI to take on higher order tasks. Ultimately, truck drivers will probably be replaced by self-drive trucks, for instance.
How long it will take to reach this level of sophistication is debatable – but it’s far from reality.
AI still faces difficult challenges. One is securing high-quality data sources for AI systems; another is how to integrate the technology with legacy IT systems. In addition, it usually takes a lot longer for cutting edge technology to spawn real-world applications than is generally assumed. Believe it or not, the first crude electric car was built in the early 1800s. Deep learning could be revolutionary, but at this time we only have a very basic understanding of the human brain’s architecture.
In the meantime, AI might actually enrich the workplace for humans. One speaker at MIT’s AI and the Future of Work conference even suggested that by taking on many routine tasks AI could herald the return of a concept that fell victim to technology a long time ago – the 40-hour working week!
This post was written by Ken Cottrill, Global Communications Consultant, MIT Center for Transportation & Logistics (firstname.lastname@example.org).