AI development will both generate and eliminate jobs. And if you missed out on the £700,000 job at Netflix, there are lots of other openings, many in unexpected fields.
Understandably, the negative aspects of AI’s impact on the workplace are the main topic of discussion. After decades of automation removing manufacturing employment, workers in the service sector are concerned about being replaced by “robots.”
However, every technological advancement both produces and eliminates jobs. Artificial intelligence is no different, at least in its current form, which generates text, audio, and video using enormous language datasets. Perhaps more unexpected is the kind of jobs it will generate.
What types of occupations are now being created?
Those with the coding and development skills to contribute to the creation of AI models or to their adaptation for specific purposes will play the most visible and evident new jobs. The UK government is aggressively seeking and attempting to encourage AI startups. DeepMind, a branch of Alphabet, the company that owns Google, was established in London and is currently mostly based here.
Not all of these positions require AI expertise. There are firms devoted to tailoring AIs for specialized jobs in specialized enterprises, such as functioning as customer service agents for certain industries or even online sales assistants, such as ChatGPT. The highest salary premiums are being offered for positions that specialize in developing and building AI models, however coding and development positions can also be found that pay very well.
Are new types of jobs being created already?
The artificial intelligence revolution is being explored by a wide range of businesses, but at this point on a lower scale than the new development roles. Netflix’s job posting for a product manager for their “machine-learning platform” has drawn particular attention, in part because the specified “compensation” is $300,000 to $900,000 (£240,000 to £710,000).
The job posting has drawn a lot of criticism in light of the writers’ and actors’ strikes, which were partially motivated by how AI would affect their own careers. Although it doesn’t specify writing or acting roles being automated specifically, it does use language like “powering innovation” and “offering personalization to members” with the “highest possible impact”.
However, not all of the new positions being produced by AI pay a high wage; in April, the UK’s NewsQuest group posted an advertisement for a “AI-powered reporter” position. In addition to being expected to meet and exceed page view targets, the new hire would work with an AI system to help write your news articles and ensure that all content produced meets legal and ethical standards. A salary of £22,000 served as compensation for all of it.
This kind of position is predicated on the notion that employing AI models effectively right now requires a skilled task. Users must keep a watch out for instances where an AI model appears to be providing real information but is actually “hallucinating” — making up false “pseudo facts” — and use extremely specific instructions rather than generic queries to get the best results.
Should everyone be considering retraining for this kind of position?
Though we don’t know, it’s definitely not a guarantee. The long-term prospects for these early AI-created roles may not be the best. According to Dr. Caitlin Bentley, a lecturer in AI education at King’s College London, “the best prospect” is not programming because coding will change, meaning that people who code AI may eventually see some of their skills displaced.
She also speculates that the new field of “prompting,” which involves giving very specific instructions to AI to produce output of higher quality, may not last very long. She also believes that prompting will soon become obsolete due to advancements in AI. Better translation and interrogation techniques are what we actually need.
So some roles created by AI are temporary?
It is undeniably true that artificial intelligence models need to be “trained” on real-world data. This information already exists naturally for language, but AI requires specific training for other applications (such recognizing traffic lights).
Clearly, these roles won’t last forever. For their autonomous cars, self-driving companies, for instance, all use “test drivers”. But many roles have comparable functions, albeit in far less obvious ways. Using reporting from the Bureau of Investigative Journalism from last year, it was made clear that Amazon employees in UK warehouses had to categorize products a certain way so that AI could recognize it.
To train the AI, workers in India and Costa Rica were then paid to confirm confusing forms all day long, frequently for hours at a time with only a few seconds for each one. The ultimate goal was for the AI to eventually replace the workers.
According to a person with knowledge of the issue, it seems surprising that two people are now engaged in a task that was previously performed by just one. But in order to assist in the training of robots that may ultimately replace both of their occupations, they are both being asked to behave like robots.
These “ghost in the machine” jobs typically involve dreary repetitive work that AIs are yet not sufficiently trustworthy at, or social media moderation for the same reasons.
Which roles are durable?
According to Alison Gow, a former senior executive at Reach (publisher of the Mirror, the Express, and other local publications), who has studied the use of AI in media firms, the areas where the technology has the biggest long-term potential are those where it is used to improve upon or alter current positions by taking on tedious chores.
Because someone would ultimately need to be in charge and lend a lens of journalistic scrutiny to generated content, she would anticipate the introduction of AI editor roles.
The positive scenario for the news industry is outlined by Gow: According to her, we’ll see more automation of utility content, like weather reports and match reports in sports, and we’ll also see AI used for greater accessibility — so transcription or translation services, or text to audio.”
She demands “real-world workable benefits for a newsroom, not doing shiny stuff for the sake of it”.
Are there any benefits in this for someone like me who isn’t particularly tech-savvy?
Businesses’ legal and human resources may be ripe for AI, according to some speculation. Gow is one among many who holds a contrary opinion, arguing that AI may need to do the exact opposite.
Businesses will require larger teams to handle the numerous challenging issues that will need to be addressed in the courts, in parliament, with unions, and elsewhere, she says, on topics ranging from copyright law to employment law and beyond.
According to Gow, the HR and legal divisions will grow to include more expertise or hire external experts on a contract basis. This will give rise to a completely new set of complaints.
Perhaps there is also hope for jobs that require a human touch. Having real human customer care representatives might become a sign of quality, comparable to how some businesses moved their call centers back to the UK as a sign of premium quality. The value of supposedly “soft skills” and service skills may rise as a result.
Whatever changes AI brings to the workplace, the people who develop the technology may not always come out on top.
