When Will AI Take Over?

January 7, 2020 • Zachary Amos


Artificial intelligence was a twinkle in the eye of
inventors and fiction writers long before our technology made the concept
viable. Now that AI has arrived in a major way, everybody wants to know: “When
will AI take over the world?” For others, the question is: “How much longer
until AI and robots put me out of work?”

The Current State of AI and Its Applications

There are too many applications for artificial
intelligence to name them all here. Compared with fallible human judgment and
cognition, AI can quickly solve complicated problems and analyze huge troves of
data to draw actionable conclusions and answer questions.

Stanford University’s “One Hundred Year Study” of artificial
intelligence, which began in 2014, is intended to be a comprehensive accounting
of AI and its global implications. The purpose of the study is to better
understand the role AI will play in various industries and the degree to which
it will displace human workers.

The study has identified the following industries as some of the ripest for AI disruption:

Transportation and freight: Autonomous vehicles will play a significant role in encouraging the adoption of AI. Driverless cars at the consumer level make streets safer, while driverless trucks deliver improved profitability and efficiency in the freight industry. In 2019, an autonomous truck made the first coast-to-coast trip in the U.S., carrying tens of thousands of pounds of butter.

Home services and deliveries: Sensors and video equipment paired with AI lead to new approaches in care for the housebound and the elderly, as well as property surveillance. Drones with AI will revolutionize last-mile deliveries and make lightning-fast turnarounds possible.

Health care: Health care service companies and hospital systems can see huge benefits from adopting AI. Artificial intelligence is increasingly capable of diagnosing illnesses with impressive accuracy and speed. It can also help spot trends or signs of trouble in historical and real-time patient health data.

Education: Human teachers will always be necessary to deliver the attention and personal touch required by academic environments. Nevertheless, AI can help students who need responsive and personalized tutoring or supplementary lessons. Natural language processing also means universities can use AI to screen students according to several factors, including their level of engagement with the admissions process.

The idea of using AI to screen students is interesting.
However, alongside the use of AI in policing and the justice system, it also
presents potential controversy.

University admissions software with AI can bring costs
down. By focusing on the objective features of a college or job application, AI
might also help remove some of the bias that’s implicit when human beings alone
serve as the gatekeepers to opportunity.

It may also amplify those biases. So far, attempts to use
AI in predicting criminal behavior and recidivism have revealed that artificial
intelligence can be just as racially problematic as human
beings. One problem is just how to define fairness in criminal justice. The
other is writing an algorithm that consistently meets whatever definition is
arrived at.

What About AI Taking Jobs Today and in the Future?

Now is the time to democratize and regulate the
development of AI rather than give in to the “black box” model. If the logic of
an AI is locked behind intellectual property laws, and isn’t even satisfactorily understood by the
people writing it, the potential for harm could well outweigh any benefits.

The other part of the puzzle when it comes to AI “taking
over the world” is the loss of human jobs. This is a worry for many individuals
in a range of professions. Is it overblown, though? One automotive company in
the Chicago area began investing in automation in 1996 and still hasn’t laid off a single employee as a result.
Instead, AI took over inspections and quality control, and helped human QA
workers pivot to higher-paying and more cognitively demanding roles.

This pattern can’t hold everywhere. The world economy is
going to shed jobs over time, perhaps for good, thanks to robotics and AI.
However, not every role is equally susceptible to robotic takeover, and AI
won’t see equal adoption in every job market at once.

The U.K.’s Office for National Statistics found that between 2011 and 2017, the number of positions at
high risk of being automated fell from 8.1% to 7.4%.

On the other hand, a recent Brookings study found that as
many as 25% of jobs in the U.S. might be at a high risk of being automated.
However, the same technologies will likely create 60 million jobs that didn’t even exist

There are reasons why Stanford’s study is called the “Hundred Year Study.” The first reason is that AI is still very much in its infancy as a technology. The second is because AI’s industrial revolution will continue for a long time to come.

So, when will AI take over? In the coming years, AI will
continue to improve business models and efficiency at every level. As AI
continues to combine with automation technologies, we must more equitably
balance both human and capital needs.