Don’t blame recruiters for searching for AI experts “with 8+ years experience in TensorFlow”. It’s what their clients are asking for.
The hype around AI brought with it a flood of experts. As literal fortune flow into the (rightfully) promising technology, thousands of job openings are advertised with titles that include “artificial intelligence”, “AI”, “machine learning”, and similar terms. Candidates with “8+ Years of experience in artificial intelligence” are offered six-figure salaries, even though until 2014 AI has rarely left the labs. AI software is sold boasting fantastical capabilities, such as “general purpose” chatbots capable of “understanding any spoken sentence”, where in reality every actual chatbot I have ever come across has trouble recognising words spoken with the slightest accent. (Good luck explaining to the automated answering service type of bots that answer phone calls to banks or utility companies that the line is bad or that you’d like to speak with a human.)
With the demand came the supply. Courses are advertised promising anyone to become a “machine learning programmer” in three months. For the right price, anyone can develop AI or sell AI. With the job offers and training courses came a wave of job candidates whose CV includes ‘AI’ or ‘machine learning’. Brief search in LinkedIn shows that there are more ‘AI experts’ as opportunities for such, with some claiming to have 5, 10, and 20 years experience. No ‘shortage of talent’, right?
Yet for those who remember, until the relatively recent success of deep neural network (2014-15) there were no such jobs such as “AI Programmers” or “Machine Learning Engineers”. Before 2015 merely mentioning “artificial intelligence” in your CV was considered career suicide even for academics, let alone professionals. Individuals who actually worked on AI before 2015 are like the legendary narwhal whales: real yet mythical, they’re hard to tell from distance, few have met them in person, and most people tend to confuse them with unicorns.
To be fair, hype is the price of every disruptive technology. The greater the disruptive potential, the greater the attention, money, and talent the technology attracts–and AI is arguably the most disruptive technology ever. Those who cannot possibly cash in their promise for AI expertise aren’t charlatans (although those exist of course) but people naive enough to believe the promise that anyone can become an AI programmer in three months. Outsiders are by definition not equipped to distinguish between signal and noise. Which is why recruiters, investors, and clients all suffer from the hype.
With AI’s success came the research grants. The new funding opportunities saw computer science departments in the UK — where money serves as a principal metric for academic promotion and demotion — enjoy surge in AI research. But along with old and new talent with genuine interest in the paradigm shift come the hitchhikers whose expertise is restricted to the latest rage in neural networks. One career administrator (ie research-inactive academia executive) who shut down the university’s reputable artificial intelligence lab (which boasted historical roots going back to Alan Turing) and completed the culling of the lab’s faculty by 2013 had a change of heart soon after and “Machine learning modelling” popped in her CV (along with IoT, Big Data, and other catchy phrases) to retroactive effect.
Today ‘AI’ is a sexy keyword, one mixed with ‘database’, ‘IoT’, and ‘blockchain’, and one which recruiters toss casually in the same breath as ‘XML’ and ‘Python’. ‘AI’ in your profile may result in being hounded by recruiters who know less about AI than they know about using search filters — seeing as the query “artificial intelligence” returns 860,000 in LinkedIn alone. But don’t blame recruiters for doing their jobs for clients who want AI experts “with 8+ years experience with programming in TensorFlow”. It’s not their fault. “Garbage in, garbage out” as programmers say.
