Teaching a machine to think like a human is the promise of artificial intelligence (AI). Using that narrow definition, it naturally follows that AI's future could ultimately include the idling of countless millions of workers who are gainfully employed today.
These concerns about job loss are logical and unavoidable, but in my opinion, they are as unfounded as they are provocative. While someday in the distant future AI systems may start to approach the holy grail of emulating the thought process and analytical capabilities of a human, today's capabilities put AI squarely in the category of a beneficial, time-saving tool rather than a human replacement.
Beginning with the Industrial Revolution and continuing into modern times, machinery has replaced workers. Automated looms disrupted the textile industry, and mass production disrupted the automobile industry. Desktop computing and word processing cut short many stenographer's careers, and other tools such as email and voicemail have imperiled letter carriers and administrative assistants.
But AI is different. AI enables the creation of work products that often cannot be replicated by any countless number of able-bodied humans. How many workers would it take to approximate the search capabilities of Google's AI algorithms? What personal shopper is so gifted that he or she could build bundles and offers with the targeted appeal that Amazon does a thousand times per second? What linguist is well-versed enough to understand the native tongue of nearly any soul on this earth?
The question is not how many people will lose their job as a result of AI, but how can people learn to use AI as a tool to improve the quality of the human existence?
Birth of the Computer Gives Rise to the Spreadsheet
Many people are surprised to learn that the word "computer" is not a new term but first appeared in the early 17th century to describe throngs of low-paid workers who toiled into the night performing repetitive calculations. This miserable task was ultimately laid to rest in the late 1970s with the invention of VisiCalc, the first modern spreadsheet. In fact, before the commercial introduction of spreadsheets for personal computers, people avoided the tedium of creating forecasts, budgets, or other computation-rich numerical models. The spreadsheet was a new tool. It made these numerical models easier to create and easier to modify. The entire concept of "what-if" analysis was essentially spawned by the advent of spreadsheets.
Did this result in large job losses? Obviously not. It made financial analysts and others more efficient and more productive. They had to learn how to use this new tool and had to evolve how they did their jobs, but in the end what the spreadsheet really did was to elevate their thinking, allowing them to spend more time on real problems, analysis, what-if scenarios, and the decisions facilitated by their results versus the manual, laborious effort of endless calculations.
The same concept applies to AI. AI is a tool that people need to learn how to use and how to apply to what they're already doing. It can improve efficiency and productivity and relieve us of some of the laborious, tedious aspects of security analytics.
And just as spreadsheets created new jobs for people who became experts at using the new tool, AI will also create new jobs — jobs focused on applying AI to security, improving AI techniques to do a better job, and on maintaining these new tools and the underlying AI technology, including tuning and data collection.
To be more precise about the new opportunities, we should separate AI-birthed jobs into two categories:
A plethora of new jobs will be created for those with expertise in applying core AI technology to new fields and applications. Experts will be needed to determine the best type of AI to use for a particular application (deep learning, expert systems, machine learning), develop and train the models, and maintain and retrain the systems as needed. In fields such as security, where vendors have empowered security software with AI, it's up to users — the security analysts — to understand the new capabilities and put them to the best possible use, just as their predecessors did with spreadsheets.
Where the AI Jobs Will Be
What AI does very well is to classify, predict, and automate repetitive tasks. Currently, for example, enterprise security analysts are faced with seemingly endless alerts about anomalous events or suspicious activities taking place within their organization. An estimated 99% of the time, these alerts are benign, false positives, according to Lastline's analysis of customer network traffic. Though I've never done it myself, I have been assured that sifting through false positives is one of the least agreeable and lowest-productivity jobs ever conceived. No product is built nor problem solved, no profit is earned, and 99% of the time one's effort is totally for naught.
Education is another well-recognized field where AI is creating new jobs. Currently, across the US, the top two positions in the list of academic openings are for security and machine learning experts. Universities simply don't have enough people and can't find professors to hire and to teach these critically important subjects. AI-powered job growth will grind to a halt and ultimately be frozen in place if higher education capacity is constrained any further.
To conclude, AI presents a tremendous opportunity for enterprising people. Employees have the opportunity to dive into a new field and abstract their job to a new, higher level of analysis and strategic value. Employers need to support these moves, allow time and flexibility, and generally stay open to employees reinventing themselves and their jobs as they embrace new technologies that will pay handsome dividends.
People won't be replaced by AI and security analysts and educators won't lose out to AI-driven security software. AI is here to stay and will be a powerful tool for improving our collective ability to defend against advanced threats. AI must be understood — its capabilities and limitations — and embraced as a new tool that already has had and will continue to have profound impact on security and other fields. But it's not something to be feared as a job killer. Nothing could be further from the truth.
Join Dark Reading LIVE for two cybersecurity summits at Interop 2019. Learn from the industry's most knowledgeable IT security experts. Check out the Interop agenda here.