AI as force for telco jobs creation is hard to imagine

Apart from the recent share price trajectory, boosted by talk of juicier dividends and future profits, Allison Kirkby hasn't altered much at BT since replacing Philip Jansen as CEO in February. The main pillars of his strategy are still in place, and they include plans to cut 55,000 of 130,000 jobs by 2030, announced by Jansen in May last year. Ten thousand have already gone, results published last week show, and the same number would be killed off by artificial intelligence (AI), said Jansen. Kirkby presumably agrees because none of the numbers has changed.

This all makes the findings of a new survey carried out by Ciena look counterintuitive. Among other things, respondents within telcos were asked this question by the optical equipment maker: "On balance, do you think AI will be a force for job creation or reduction within service provider businesses?" Despite BT's plans, and societal fears about job-stealing robots, some two thirds of respondents answered creation.

Wishful thinking? Telcos are an odd bunch, forever grumbling about regulation and Big Tech freeloaders while gibbering in excitement whenever a new technology pops up. The Mobile World Congress (MWC) tradeshow in Barcelona each February is a good place to observe this manic enthusiasm. But it is all in contrast to the industry's performance over the last ten years. Sales growth remains elusive, costs are high and the gap between techcos and telcos – in terms of market capitalization – looks galactic.

The telco response to this has involved a sharp focus on efficiency and cost cutting, and the combined headcount of the world's biggest telcos in North America and Europe has plunged accordingly. Between 2015 and 2022, 20 of the largest operators tracked by Light Reading slashed about 384,000 jobs, more than a fifth of the total. No doubt, outsourcing and divestments account for a decent chunk of that figure, but the trajectory still looks much like the average European telco's share price – sloping down.

A bubble set to explode?

Against this backdrop, AI risks being the latest technological trinket whose shine will eventually fade. Unfair? AI is undoubtedly on a bigger scale than network specific technologies such as SDN, NFV and 5G – all of which have been through their own hype cycles without delivering the payoff – and affects multiple industries besides telecom. But that means this bubble could burst with a spectacular bang.

And, make no mistake, experts are calling it a bubble. Commenting on Big Tech results in a recent blog, Richard Windsor, an analyst with Radio Free Mobile, noted the danger to Microsoft's lofty valuation "should the AI bubble burst, which it surely will at some point." As recently reported by the Economist, Alphabet, Amazon, Meta and Microsoft have collectively promised to invest more than $200 billion this year in data centers, chips and other AI-related gubbins. That would be a 45% increase on spending last year, said the Economist, using the words "overhype" and "overbuild" in its story.

For telcos, there are several problems with the notion AI will create jobs. The first, not telco specific, is simply that coffee-shop and pub talk among people outside AI and software circles is generally at odds with these sorts of survey findings. Even individuals taking advantage of tools like ChatGPT are worried about what more advanced AI will mean for their livelihoods. Anyone genuinely excited is a rare breed.

This is backed up by another survey of both ordinary people and so-called experts carried out by North Carolina's Elon University. The findings, published earlier this year, showed that 55% of ordinary people fear AI will have a negative impact on jobs, versus only 43% of experts.

They could be wrong, of course, and people in the 2020s seem prone to worrying, a nasty habit the recent pandemic probably didn't help to address. Sea levels are rising, political extremism is out of control, WWIII is imminent, we aren't prepared for the next pandemic. AI fears just fit into this catalogue of modern-day anxieties. And despite a social media-fueled perception that the world is going to hell, various data show it is a far safer place in the twenty-first century than it was in any other.

Unfortunately, for telco employees, the instincts of bosses like BT's Kirkby – under pressure from shareholders – will be to continue slashing costs unless sales are on the rise. One of the targets she announced in her first quarter update last week was to reduce annual expenses by another £3 billion (US$3.8 billion) in the next few years. Eliminating 45,000 jobs should obviously help.

AI will only buck this downsizing trend if it spurs sales growth. Otherwise, the inclination of company bosses will be to examine where it could save them money, and that won't happen by fattening the workforce. The whole problem with the AI story right now is that it isn't generating sales growth for anyone except a few chipmakers (mainly Nvidia), a cluster of Big Tech firms, and the companies that help build and connect data centers. Ciena, unsurprisingly, is among them.

Making money from AI

The respondents to Ciena's survey are nothing if not optimistic. "How confident, if at all, are you that CSPs [communication service providers] will be able to monetize AI traffic over their networks?" they were asked, to which 85% of them apparently said either very or somewhat confident. Money will accrue from such things as "opening your network to third-party GenAI integrations," "security and privacy services," "new product offerings" and "tailored subscription packages," said the survey results.

But judging by the depressed share prices of telcos in the Western world, investors do not believe it. And why should they? Telcos have never been able to generate meaningful revenues from anything besides basic connectivity – whether for calls, text messages or web browsing – and monthly average revenue per user (ARPU) has been skydiving for years. Among Deutsche Telekom's postpaid mobile customers in Germany it has descended €4 in the last decade, to €19 for the recent first quarter of 2024.

This is not to say AI won't create any jobs at all within telcos. To exploit AI, telcos will need expertise in various fields, said respondents to Ciena's survey, with "cybersecurity," "machine learning" and "programming and coding" voted as the top three. But not everyone is suited to these computer science-type jobs, and no telco will need as many people in software-related roles as it traditionally employed in customer service and other departments susceptible to automation.

What further contradicts the view of AI as a creator of jobs is telco talk of "zero touch" and "autonomous" networks, manageable without human intervention thanks to advanced AI. One could feasibly imagine a basement of sweaty coders manipulating the AI that controls the network. But this sounds neither zero touch nor autonomous – more like the armies of programmers reportedly hidden behind certain chatbots, ready to zap any interfering hallucination.

A different scenario is that AI does most of the coding. Software that writes other software already exists in the form of technologies such as Microsoft-backed GitHub Copilot, which supposedly takes the drudgery out of programming so that humans can focus on the more intellectual and valuable stuff.

But there might not be much valuable stuff left. "Eighty percent to 90% of any code you write could be handled by AI today," said Tareq Amin, the CEO of Aramco Digital, when he recently met Light Reading at this year's MWC. "And we're on a path where the AI will write much more efficient and cleaner code." Still feeling optimistic?