Learning Lab #4 Part I: Fear not the AI

Artificial Intelligence (AI) is creating a tremendous buzz in the tech industry and far beyond, yet it seems to inspire fear rather than excitement for many. The concerns are generally of two types:

1. The machines are taking over the world!
2. My job will be replaced by a machine!

The first is undoubtedly a result of decades of dark plots in literature and cinematic narratives such as “Ex Machina,” “The Terminator,” “Transcendence” and the like. Although we can safely leave this anxiety in the capable hands of science-fiction writers, the second fear is more grounded; AI does have some potential to disrupt many jobs that exist today. That said, it’s also true that those willing to approach their work with AI in mind may find tremendous opportunity to do and be more than they otherwise could.

To consider how the tax office might thrive with the introduction of AI, it’s helpful to consider some definitions. AI is an extremely vague term. Like “the internet” and “the cloud,” it represents a lot of different things to different people. And those concepts vary depending on what function AI is performing; as a panelist noted at the recent PHORUM 2017 technology conference in Philadelphia, “AI is only AI until it does something. Then we call it something else.” For instance, if we can talk to an AI and it responds to us (e.g. Siri), then it’s natural language processing (NLP). If an AI is looking at photos or videos to identify characteristics or patterns, it’s image recognition. If it’s reading and interpreting text from an image of a form or document, it’s optical character recognition (OCR).

With that principle in mind, here are definitions of three functional areas of AI worthy of attention in the tax office:

1. Cognitive Computing (CC): Forbes contributor Bernard Marr describes CC as “simulated human thought processes in a computerized model. Applying self-learning algorithms that use data mining, pattern recognition and natural language processing, the computer can mimic the way the human brain works.” CC could help the tax office to identify data patterns that point to both risk areas and opportunities.
2. Machine Learning (ML): Expertsystem.com describes ML as “systems [with] the ability to automatically learn and improve from experience without being explicitly programmed. The process of learning begins with observations or data…in order to look for patterns in data and make better decisions in the future based on the examples that we provide.” ML can help tax departments build on previous learning to support better decision making.
3. Robotic Process Automation (RPA): An article on techtarget.com defines RPA as “software that automates other software, performs high-volume, manual, repeatable tasks by mimicking how humans interact with computers…and tak[ing] on human tasks that call for problem-solving skills and judgment.” RPA could enable tax departments to bypass the mundane, repetitious, urgent work that diverts talented tax minds away from opportunities to better advise and protect their business stakeholders.

No need to administer the Turing test to determine whether these systems have “achieved consciousness,” and no need to fret about the rise of “The Machine!” Rather, these technologies should be viewed as forms of “AI-assist” that can do much to eliminate low-value activities through algorithms and automation, freeing up tax executives and professionals to focus on more advanced tax strategy and planning.