When Deloitte asked leaders in the HR field what they thought AI would bring to their businesses, they found that 38% of respondents had begun the transition to AI already.
Sixty-eight percent expected to begin implementing it by 2018. This isn’t going to have an abstract effect: we’re all going to go to work with the results of the AI revolution. Many of us already do. So which job roles and tasks are going to change the most, and how will that affect the tech sector on a day-to-day basis?
These 5 jobs will be changed utterly by 2022 as AI remakes the workplace.
1: Software Developers
We’ve all had time to get used to the idea that AIs can beat human game players, even in complex games like chess and Go. But what’s more unexpected is the possibility that AIs might move on to creating games, or augmenting human developers as they create games, apps and even full-blown enterprise platforms.
Creative, or pseudocreative, AIs have a long pedigree. We can look back to Simon Colton’s AI experiment, the Painting Fool, ‘a computer program and aspiring painter.’ That was interesting – but ANGELINA is the AI to make developers look up. Standing for ‘A Novel Game-Evolving Labrat I’ve Named Angelina,’ the project is the work of Michael Cook, researcher and PhD student who hopes it’s a step on the way to software that can interact with human culture.
It’s familiar in AI to see replacement technology – machines that can paint, write, design, or develop – hit the headlines. These projects won’t be in the workplace for a while, if ever. But developers will likely see the necessity to hand code gradually disappear. That’s part of a trend that’s been going on for a while – once, you had to be able to code to a decent level to get anything out of a computer. No HTML, no website. Now, graphical interfaces increasingly merge design and development.
The restrictive graphical interfaces nondevelopers use to design and develop, essentially mixing and matching a range of off-the-peg elements, will be replaced by AI-augmented interfaces that autocode for requested items – potentially replacing or disrupting the lucrative market for WordPress templates among other businesses.
Where the developer at work now hand codes or checks GitHub when she isn’t sure, these processes may both be automated, drawing on the web and internal collections of previously coded solutions initially.
What effect will this have on development more broadly? After all, it’s developers who program the AI, right? Those jobs should still be safe.
Actually, that’s not something to rely on. AI is becoming quite productive of novel code. Google’s AI-enabled translation software is generating its own internal language; Facebook’s AI experiments created their own language and began talking to each other before anyone managed to shut them off. But those pale beside Microsoft’s DeepCoder, which presages an age when AI writes code instead of following it.
Five years out, don’t be surprised to see development as an ‘AI-supervising’ role, and development as a standalone discipline disappearing as a business skill.
As that happens, other roles in software engineering will be affected too. Generally speaking, the more a role deals with judgement and foresight and the fewer repeatable, structured actions it requires, the later AI is likely to affect that role. Architects, for instance, will face serious competition from AI later than software developers. And quality assurance might get some AI augmentation, but the human element of crafting software for humans to use means they’ll be in their seats for a while yet.
As CIOs struggle to accommodate the changes in their role that AI brings, there’s been the suggestion that the ideal solution is to give AI its own department and name a CAIO – a Chief Artificial Intelligence Officer – to the C suite.
Head of sales and development at Rainbird, Matthew Buskell, says:
I would argue, however, that AI presents opportunities in strategic Innovation like Digital Transformation and Industry 4.0 but it also dramatically impacts process improvement projects found across the business. Therefore there is merit in having this as a role reporting directly into the CEO.
At the end of the day, the fact is AI is so profound it deserves its own leader and department.
Writing for HBR, Andrew Ng concurs: ‘To the majority of companies that have data but lack deep AI knowledge,’ he says, ‘I recommend hiring a chief AI officer or a VP of AI,’ adding that ‘some chief data officers and forward-thinking CIOs are effectively taking on this role.’
This change isn’t by any means certain, and in March this year HBR also ran a piece by Kristian J Hammond, AI research scientist at the MocCormick School of Engineering at Northwestern, entitled, ‘Please don’t hire a Chief Artificial Intelligence Officer.’
‘In much the same way that the rise of Big Data led to the Data Scientist craze,’ argues Hammond, ‘the argument is that every organization now needs to hire a C-Level officer who will drive the company’s AI strategy.’
But simply having an AI strategy isn’t enough, Hammond argues: instead AI needs to be integrated into the business in the service of business goals, not given its own department.
The counter-argument comes from Andrew Ng who says, ‘the benefit of a chief AI officer is having someone who can make sure AI gets applied across silos.’
Another argument against the creation of the role is that in a few short years, the distinction between ‘AI’ and ‘non-AI’ may have begun to fray completely. At that point, though, the CAIO role could begin to fade over into marketing and customer success.
The final distinction between a business that gets a CAIO and one that doesn’t might be the one suggested to SearchCIO by Chrag Dekate, research director at Gartner: ‘one rule of thumb to identify whether an organization needs a centralized function is to decide whether its core business model is likely going to require a fundamental change. If the answer is “yes”, then a chief AI officer is probably needed.’
Assuming that as well as information and technology officers, AI gets its own seat in the C suite, what will that look like?
The CAIO’s role in the organization will be less to be a subject matter expert than a generalist who can explain AI to the rest of the C suite, and implement it in constructive ways – for instance, marrying AI’s possibilities to the ‘intrapreneurial’ hothousing of potentially disruptive innovations within already-established businesses. If there’s no CAIO to do this, it’s likely to become the joint responsibility of the CIO, CTO and CEO.
3: Data analysts
At first glance, data analyst roles are among the most vulnerable to AI-driven change. It’s a role that’s about crunching and analysing data: surely AI is intrinsically better at that than humans?
After all, ‘AI can handle and organize an amount of data that is beyond anything a human could hope to consume or comprehend,’ in the words of Dominic Namnath, CIO at Tri-Counties Regional Center, a nonprofit in Santa Barbara, CA.
Plenty of businesses are likely to start out with this view too. It’s the data analyst’s job to get the organization to understand that the business isn’t going to be able to hop over the implementation of best data analysis practices and land in an AI-enabled paradise of data clarity.
In fact, one of the prime markers that shows a company is ready for AI implementation is that it has mature, effective data analysis in place already.
When AI does come into the data analyst’s role, it will be as a new tool and then as a subordinate co-worker, rather than as a replacement.
Data analysts will find that only after they have created structures that enable them to automate basic data flows. If that isn’t done, adding AI further upstream won’t improve the efficacy of the data analyst or the quality of the information that underpins business decisions, because the AI will be working with out-of-date information.
However, some businesses will see AI transform the processes by which data is collected. For instance, manually-entered CRM data is notoriously incomplete and unreliable. But as major CRM providers embrace AI add-ons or bake AI into their offerings, this data is likely to become more reliable and require less collation since it will be ported across from one digital source to another automatically.
But ‘without basic automation’, warn Nick Harrison and Deborah O’Neill, ‘strategic visions of solving complex problems at the touch of a button remain elusive.’
The role of the data analyst will increasingly involve manipulating and designing structures to allow AI to do the work of actually analysing the company’s data. But this will simply accelerate the chaos without data analysts who can structure information flows that centralize and distribute vital data efficiently.
Most management jobs will survive, though many will be radically changed as tasks and whole work types are automated.
As simpler, repetitive tasks are automated, rapid decision making and judgement calls will move to the fore.
Within the next five years, a manager’s workday will shed its admin load and much of the meeting time. Instead time will be devoted to strategic and tactical thinking, judgement and crucial people skills. Midlevel management will be the tier most sharply affected by these changes, while C suite operatives will see major but relatively smaller changes.
The average manager is essentially working two jobs. One is focused on their immediate job goals, while the other is aimed at maintaining their portion of the company data lake, handling the admin for their staffs and their departments.
That’s going to change.
AI will radically change this time and effort allocation in two ways:
1 – Much project management and time management will be automated
2 – Much manager-level admin will be bot- or AI-assisted.
Some of these will make admin tasks self-service. For example, if a member of staff wants to book holiday, they might now ask managers directly or message them. Instead, an automated system can let staff enter the dates they want and find out if someone’s available to cover that time. For example, Subzz already offers a lot of this functionality. How long before something like this becomes the norm?
Many managers spend a lot of time on project management tasks – essentially, assigning sections of work to different staff, then checking that they’re doing that work. This is a classic opportunity for AI to step in and automate nearly an entire role: for most managers, project management duties will all but vanish from their to-do lists. (We’ve also seen some more basic roles disappear completely in some businesses – Fukoku Mutual Life Insurance replaced 30 roles with AI in January this year.)
As more effective AI assistants reach market, managers will rely on bots to tell them what’s next, alert them to vital incoming messages, automate their appointment setting and reply to common inquiries from superiors and staff.
HR has always involved making complex decisions based on incomplete data. Writing for IBM, Blythe Howard-Chou says that recruiters on average would not rehire 39% of their recent hires. Yet Howard-Chou goes on to observe that just 7% of companies use analytics to predict and achieve hiring success.
IBM’s own research indicates that this is going to change, significantly and soon. Forty-six percent of respondents believe AI will transform their talent acquisition. But what will that look like on the ground?
Automated screening and reduced bias
Biased hiring harms companies in several ways. Apart from the reputational damage, which no-one can afford, it robs the business of the best candidate for the job. By automating screening, AI will offer HR the opportunity to accelerate the early stages of hiring by immediately removing unsuitable candidates below a certain threshold. But after this stage has been completed, AI steps in again to ensure that a name, zip code or other irrelevant data point doesn’t stand between the right person and the right job.
Augmented onboarding and training
Onboarding approaches differ hugely between consumer and employee. Employees are typically onboarded by processes that take involvement for granted. They then go on to cost the company thousands of dollars in lost productivity and data risks thanks to noncompliance with protocols they only half remember, and nonadoption of tools and systems they’ve forgotten how to use. Security breaches due to employee noncompliance and error average $3.85 million. Gallup says 87% of employees are disengaged at work.
With consumers, we don’t take this approach. Consumer-facing mobile apps that cost a few dollars put more effort into incentivizing onboarding than businesses making midcareer hires. AI will change that quickly, offering comprehensive onboarding programs that move employees more quickly and effectively towards full productivity by borrowing techniques from gamification and consumer onboarding more than traditional employee procedures.
Significantly, there’s actually scope for full replacement of some low-level HR operatives within the next five years. The National Bureau of Economic Research reported this year that AIs were more effective at directly making correct hires than humans.
AI will change the way we work in a way we haven’t seen since the nineteenth century. We’ll see job roles that used to be labor-intensive switch over to relying on robots to do the actual work – just as we did with physical productivity in factories in the late twentieth century. That’s going to see coders and developers do more supervising and strategic work, less day-to-day productivity. And that pattern will be repeated across the organization, particularly affecting management and data professionals.
By Eugene Rudenko
The race to net-zero emissions will forever change the way many companies do business. The immediacy, pace, and extent of change are still widely underestimated. Early movers can seize significant advantage. In this report, coauthored with the WEF Alliance of CEO Climate Leaders, authors explore how other companies can take a similar path by identifying, creating, and scaling green businesses.
The current debate over ESG and sustainable investing is noisy and sometimes rancorous, and the temptation is strong to just tune it out until it’s better resolved. But, in the end, leaders must resist this urge and accept that it’s a relevant discussion.
Looking at today’s consumption levels, sustaining our current growth trajectory would require the ecological resources of 2.3 planets by 2050. This number is significantly higher for mature markets. The US, for example, would need five planets to sustain present-day consumption levels; Germany would need three.