AI’s disruptive reputation belies its potential for positive change
Leads Thomson Reuters’ Toronto-based Centre for AI and Cognitive Computing and is head of its global corporate R&D team
Artificial intelligence (AI) and machine learning are already changing how we work, how we shop and how we connect with each other. But their “disruptive” potential has caused much concern. A report by McKinsey Global Institute estimates that in about 60 per cent of occupations, one-third of the tasks could be automated – globally. The report states that AI can transform some business activities and has the potential to fundamentally change others. However, the AI story is just beginning to unfold and we do not yet fully comprehend the potential and/ or impact of these technologies.
This impact, to the extent that it materializes, is massive in terms of how work gets done and what new jobs will be created in the AI economy. This opportunity is driving global investments in AI, and governments here in Canada are funding various AI initiatives such as the Vector Institute and its Supercluster Initiative. The lead private-sector funders of Vector represent various sectors including the banks, retailers, manufacturers and information companies. Concurrently, businesses are racing to create their own AI teams.
As practitioners who build AIenabled applications, like most others our objective has never been to disrupt how work is done, rather to remove customer painpoints, to analyze complex tasks and ask how to achieve more with less while delivering an intuitive user experience. With this in mind, we see three broad types of opportunities for artificial intelligence.
Content automation: AI is already driving scale and automation in how information providers collect and aggregate content, how it’s enhanced, organized and delivered. Most creative work (content authoring) remains a manual and time-consuming exercise, but “functional writing,” such as contracts, will likely be automated.
Digital business: Digital business is the transformation of business activities and making them data-driven. Examples of where machine learning is heavily utilized include business intelligence and advertising. The future where companies can personalize their marketing programs to individual shoppers through their loyalty programs is closer than you think. Once retailers “know” their customers they will be able to use dynamic pricing and personalized marketing to achieve maximum returns. The transformative potential of AI with regard to supply chain management and forecasting is intriguing. Imagine the efficiencies achieved when manufacturers can accurately predict demand.
Enabling knowledge work: At a high level, knowledge workers do three things: (1) they search for information. AI is already heavily utilized in “find” technologies such as search engines. The challenge will be to enable machines to “learn” from a series of questions and to deliver intuitive context-preserving experiences while tackling information overload. (2) As workers find information they try to analyze it. Tools already exist to analyze and process data, but the real opportunity is the evolution from statistical analysis to deeper understanding of documents and data sets. Finally, (3) knowledge workers make decisions. To date, AI has been limited to playing a decision-support role. While likely a good thing, there are many instances where real-time requirements do not allow a human in the loop, such as fraud detection.
The opportunities are significant. However, in the near term we should expect broad yet incremental change. Boston Consulting Group surveyed more than 3,000 companies last year and found that while 85 per cent of them believed AI would become a competitive advantage in the future, only a quarter were implementing it now, and only 5 per cent were implementing it extensively.
What we do know is that big data, computing power and connectivity are changing the industrial landscape. The volume and diversity of information continues to increase. At Thomson Reuters, we process more data in a single day than we did in a month just five years ago. The explosion of information is increasing the demand for automation and AI. The opportunity rests in accelerating the digitization of businesses, making them more data driven by building applications that deliver machine-assisted insights.
We also know AI works best when combined with human expertise, to augment us, not to replace us. This requires intuitive, task-focused and user-centric AI applications and workers trained in how to use AI technology.
Will AI disrupt the work force? Change, yes. Disrupt? Not yet. Yes, sometime in the future, when driverless cars are commonplace, there will be no need for taxi and bus drivers – but until then, the opportunity and the challenge for companies will be to identify where AI can solve the real problems of today that offer a competitive advantage. But the technology will only be part of the solution. The extent that we are ready for the change and enable our workers to thrive through this change may be the bigger challenge.
Big data, computing power and connectivity are changing the industrial landscape.