As previously published on 6/28/23 in MarTech
By Scott Gillum
Estimated read time: 5 Minutes
What if they are wrong?
When responding to questions about AI replacing humans in certain roles, most ‘experts’ claim that AI will replace some jobs, but will be a much more valuable tool for augmenting human intelligence and ability.
In all of the hype associated with this latest technology wave, an important trend is occurring across industries that could significantly change the impact of AI – the retirement of the knowledge worker.
We need to look no further than the last wave of intelligent technology – the “internet of things” (IoT) to see the impact.
The term ‘Internet of Things’ was coined in 1999 by computer scientist, Kevin Ashton. While working at Procter & Gamble, Ashton proposed putting radio-frequency identification (RFID) chips on products to track them through a supply chain.
“Machines talking to machines” started rolling out in early/ mid 2010 making their way into manufacturing, precision agriculture, complex information networks, and for consumers in a new wave of wearables.
Now, having about a decade of experience of how IoT has impacted certain industries and markets, perhaps it can give us some interesting insights on the future of AI.
In 2010, Cisco launched the “Tomorrow Starts Here” IoT campaign at the time when communication networks were transitioning from hardware “stacks” to software development networks (SDN).
The change meant that in order for carriers to expand their bandwidth, they no longer needed to “rip and replace ” hardware. They only needed to upgrade the software. This transition began the era of machines monitoring their performance and communicating with each other, with the promise of one day producing self healing networks.
Over this same period, network engineers who ushered in the transition from an analog to digital began retiring. These experienced knowledge workers are often being replaced by technicians who understand the monitoring tools, but not necessarily, how the network works.
Over the last dozen years networks have grown in complexity to include cellular, and the number of connections has grown exponentially. To help manage this complexity, numerous monitoring tools have been developed and implemented.
The people on the other end reading the alerts see the obvious, but have a difficult time interpreting the issue, or what to prioritize. The reason is, the tool knows there is an issue but is not smart enough yet to know how to fix it or if it will take care of itself. Technicians end up chasing “ghost tickets,” alerts that have resolved themselves, resulting in lost productivity.
The same thing is repeating itself in marketing today. As one CMO told me; “I can find people who know the technologies all day long, but what I can’t find is someone who thinks strategically. Ask a marketing manager to set up the tools and run a campaign and they have no problem, but ask them to write a compelling value proposition or offer for the campaign, and they will struggle.”
It’s easy to get sucked into the tools. AI generators are really intriguing and can do some amazing things. But based on what we have seen, the tools are not smart enough to fully deliver on their promise…yet.
Here’s the warning from IoT – as tools become more knowledgeable, the workforce operating them is becoming less. It is leaving a knowledge gap. As that knowledge is transferred from worker to machine, we need to ask ourselves what we’ll be left with. Will there be enough experience and expertise in our workers to know if what comes out of the machine is accurate, factitious, or even dangerous.
In a recent WSJ article, Melissa Beebe, an oncology nurse, commented on how she relies on her observation skills to make life-or-death decisions. When an alert said her patient in the oncology unit of UC Davis Medical Center had sepsis, she was sure the AI tool monitoring the patient was wrong.
“I’ve been working with cancer patients for 15 years so I know a septic patient when I see one,” she said. “I knew this patient wasn’t septic.”
The alert correlates elevated white blood cell count with septic infection. It didn’t take into account that this particular patient had leukemia, which can cause similar blood counts. The algorithm, which was based on artificial intelligence, triggers the alert when it detects patterns that match previous patients with sepsis.
Unfortunately, hospital rules require nurses to follow protocols when a patient is flagged for sepsis. Beebe could override the AI model, if she gets doctor approval, but faces disciplinary action if she’s wrong. It’s easy to see the danger of removing human intelligence in this case, it also illustrates the risk associated with over relying on artificial intelligence.
AI will free us from low value tasks, and that is a good thing, but we need to redistribute that time to better developing our people, and our teams. The greatest benefit from these game changing technologies in the business to business environment will be realized when we combine equal amounts of human intelligence with machine intelligence.
