You are currently viewing The Human Impact of Artificial Intelligence and Change Management

The Human Impact of Artificial Intelligence and Change Management

A few weeks ago, we discussed some of the benefits and challenges facing the mass adoption of new artificial intelligence technologies for companies, governments, and society as a whole. Today, we want to focus on the human impact of artificial intelligence and how these effects can be minimized through timely global governance (and common sense).

Unfortunately, history has often shown a tendency for people and businesses to adopt a new technology or innovation without truly understanding the ramifications. From the industrial revolution, to the internet, to social media, to cloud networking, the impact of technological revolution goes beyond the economic and the political – it affects all of us, often in unexpected ways.

While we are as excited as anyone else about the opportunities this new generation of AI technology offers, we believe it’s important to put people first and place the needs of our communities over those of big business.

The Impact of Automation on the Labour Market

Just as robotic assembly lines and supply chain automation resulted in the elimination of thousands of jobs in the manufacturing sector, within the next 25 years generative AI tools threaten to have the same effect on office workers, doctors, lawyers, and teachers as robotic automation did to factory workers. 

While this will result in conveniences for consumers like passing through airport customs with their fingerprints and facial recognition or balancing their stock portfolio with real-time market analysis tools, what percentage of those very same consumers will lose their livelihoods thanks to the very same technology that makes these conveniences possible?

In most developed countries, the service sector makes up a significant portion of the total labour market – around 80% for the United States alone. These are many of the same roles that are at the highest risk of AI automation, not simply fast food workers or customer support jobs but refined roles in complex industries ranging from healthcare, to finance, to marketing. 

Possible Solutions: This is a difficult area to analyze as it’s a global issue that countries are trying to address in isolation, and in a global market single actors risk becoming uncompetitive and falling behind. For instance, many countries are experimenting with instituting a 4-day work week, but such a culture shift would require global alignment and cooperation or else it’s doomed to failure. 

Another option that has been discussed is implementing a universal basic income (UBI) to combat the inevitable job losses resulting from automation. This would not only ensure a basic standard of living for all citizens but also allow for a realignment of the social safety net from maintaining a variety of expensive and inefficient programs to providing people utility directly.

However, when considering the social impact of such measures we need to ask ourselves, do people want to be idle and without work? Would they be happy without the personal satisfaction of being productive? What about the increased tax burden on an already strained middle-class?

A less radical solution would be providing employment subsidies to businesses that employ a certain number of human workers. While this would validate complaints that many offices are little more than “adult daycares,” it would be an easy and more cost-effective method to insure our accustomed standard of living will survive in the AI-driven economy of the next decade. 

Democratic Impact

“Fake news” and misinformation have become powerful weapons for hostile governments to spread discontent among their enemies and to manipulate their own citizens. How much worse will this get as deepfake videos and AI voice generators are able to create videos of world leaders indistinguishable from reality? Already we see how easy it is for blatant lies to become accepted as truth so long as enough talking heads in the media repeat it.

Likewise, as moderation of hateful content becomes more and more automated, do we truly want to leave it in the hands of private enterprise to determine what counts as free speech and what should be censored? Particularly as content moderation and enforcement becomes more automated, it will become all the more important to ensure free expression online enjoys the same rigorous legal protections it does offline.  

For instance, what avenues of recourse should smaller creators have when their content is flagged as inappropriate and removed by content moderation algorithms? So many social platforms like Reddit or YouTube are built on user-generated content, but they are the first to declaim moral and legal responsibility to the people who create content for them, making it difficult to appeal when they are unfairly targeted by blanket content moderation policies.

Put simply, it should not require rallying a fan-base and using threats of legal recourse in order to speak to a representative of the company about why their content was removed or why their channel was banned. Small- and medium-sized content creators put in hundreds to build a following and a livelihood and all that can be erased in a single arbitrary sweep. 

Possible Solutions: Quite simply, as AI video, audio, and image generators become more refined, companies need to provide a way to quickly and easily tell if content is real or generated, whether that be a watermark or a marker within the file’s metadata. While talented hackers would certainly be able to find ways to bypass such protections, having a baseline level of trust and disclosure would go far when it comes to fighting the capability of AI for disinformation – especially when it comes to crimes like AI revenge porn and fraud.

Likewise, AI content should be clearly labeled on social media and other platforms, but as creators experiment with integrating AI into the work there must be a path to appeal blanket content moderation policies.

Customer Support Headaches

At one point or another, I’m sure we’ve all struggled to get through automated customer support lines and chatbots to try to speak to a flesh-and-blood human who can understand nuanced issues and provide the support we need.

While these chatbots  and automated virtual assistants will become more and more refined as they become integrated with large language models, sometimes you just need to talk directly to a person to get your point across. Especially when it involves serious issues such as your medical information, investments or finances.  

Unfortunately, the integration of AI into customer support will eliminate more and more human operators as companies rush to reduce their operating costs, increasing the runaround you will have to put up with before being directed to a human. Not only will this cost millions of jobs, but it will take a minor inconvenience everyone faces and worsen it ten-fold. 

Possible Solution: As frustration with fully-automated customer support deepens, offering a staffed customer support line may become a valuable differentiator and competitive edge for businesses that can support it. Government regulation could also play a role in ensuring businesses maintained certain standards around customer support in order to make sure people who require critical aid  get the help they need in a timely manner, especially in fields like health and finance. 

Managing the Global Impact of AI

AI itself needs to be discussed in terms of social and human impact. What is the point of technological and economic progress if it does not make our lives better? As history has proven, we will get there eventually, but what is the short-term impact and how much suffering is acceptable on the way there? How will global cooperation be ensured and managed amid this monumental shift? 

Look at the loss of manufacturing jobs in Canada following the establishment of the North American Free Trade Agreement (NAFTA) in 1992. Hamilton alone lost over 30,000 blue collar jobs, people with over 25 years of experience lost their livelihoods and their way of life all at once. Even now the impact is still being felt in towns and cities across Canada that lost their main source of employment. 

Canadian lawmakers at the time fully understood what the outcome would be and decided that it was an acceptable sacrifice, they introduced the Goods and Services Tax (GST) in acknowledgment of the shift of the tax base away from a manufacturing economy to a service economy.

Today, politicians and business leaders have the opportunity to avoid making the mistakes of the past by placing the good of communities and societies ahead of quarterly growth – however that’s a difficult task in a globalized economy with global competition. That’s why a global mindset is required to address the variety of challenges presented by AI automation.

As discussed above, this is a global issue that will require global governance to address. Whether that’s through treaties or establishing independent agencies geared towards change management, it will be a test of global cooperation with major economic and social ramifications for every country on this planet.