TECH TALK: Algorithms are a key feature of data analytics, but the metrics don’t always make sense, says BILL MAGEE
Those pesky artificially intelligent algorithms are at it again as bot-by-bot they stealthily take over our very human, but rather chaotic, business lives. Computerised automated analytic/predictive software programs have already been blamed for sending a self-driving car into a crash zone and wrongly downgrading thousands of students’ exam results. Now, a Harvard Business Review (HBR) study concludes that AI is bamboozling talent-hungry businesses in their efforts to fill vital skills gaps.
No doubt, representatives from some of those very companies listened intently to an impressive line-up of speakers at last week’s Data Summit. First Minister Nicola Sturgeon promised the Edinburgh gathering that ethical use of data and AI is central to her government’s policy.
Yet, uncertainty lingers. When 189 delegates were asked if they thought Scotland is on track to become a leader in ethical and inclusive AI, just 57% said “yes”. Not an overwhelming margin by any means.
HBR highlights management and leadership challenges amid the “buzz” and “hype” surrounding AI problem-solving. Typically, algorithms are employed through what’s known as “means-end analysis” towards finding a path to a target goal.
In an (unreported) recruitment report of organisations based in the US, UK and Germany, nine-out-of-ten surveyed admit their experience in using internet algorithmic (ro)bots (or is it robotic algorithms?) is apparently resulting in qualified and highly-skilled candidates being “vetted out” of the process.
It seems recruiters are being repeatedly misled by such inflexibly configured metrics giving feedback that is plain faulty.
Digital algorithms conjure up images of “The Matrix” and “Robocop”, highlighting a time where AI takes over all our jobs, pushing humans to the sidelines. If only we hadn’t given technology so much free rein.
Author Cathy O’Neil ratchets up the uncertainty in her book “Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy”.
The academic warns against blindly trusting algorithms when making sensitive decisions claiming they can prove harmful to many people.
O’Neil cites “algorithmic bias” along with dangerous feedback loops expanding into all sectors of daily life including the economy and criminal justice system.
HBR agrees, claiming such AI-driven systems are letting organisations down, at the very time when they are in desperate need to tap talent pools.
Computerised sets of mathematical instructions set to supposedly solve a problem regularly, instead, apply over-strict conditions that exclude perfectly viable candidates for a job.
HBR’s survey reveals that 88% of employers admit qualified and highly-skilled individuals don’t make it in the recruitment process.
The study found that by ill-matching CVs to commercial needs, the individual fails to match the exact criteria established by a job description as mapped out by the algorithm.
Such automated systems are increasingly representing the “foundation of the hiring process” with 90% of employers surveyed also saying they used such systems because they believed they will “initially filter or rank potential middle-skills and high-skills candidates”.
HBR says this is now shown to be a mistake. Yet, if handled properly, algorithms can and do prove essential to business and commerce.
A prime example is Barracuda Networks, a Gartner 2021 Magic Quadrant “Visionary” and subject of a joint webinar with Capito IT Services and Solutions.
Barracuda’s leading spam firewall proposition utilises several algorithmic filters to ensure businesses, people and data are secured and protected.
HBR contends the same digital standard should apply to the skills field.
An algorithm can successfully form the nucleus of a job description and requirements, but its filters need to be shifted from “negative” to “affirmative” in the automated recruiting process.
Only when endowed with such new metrics can “talent” be better evaluated and organisations gradually fill that growing skills gap.
Or, better still, in the final analysis actually see all the applicants in person. At least via Zoom? Just a thought..
- Capito won the “Best Cyber Breakthrough” prize earlier this month at the Scottish Cyber Awards 2021