Talent, the attribution someone makes about our ability to excel in a socially valued task or job, has been a sought-after commodity for centuries. A documented history of formal talent assessments dates back to the Han dynasty in China (202 BC – 220 AD). In modern times, talent has occupied a central place in corporate strategies, courtesy of McKinsey’s famous “war for talent” notion, which postulated that the main competitive advantage of organizations would fundamentally rest on their ability to find, grow, and retain talented people.
A lot has happened in the world since then: e.g., the dot-com boom, 9/11, the global financial crisis, the meteoric rise of Big Tech, and of course the current global pandemic. Yet a constant feature throughout this time has been organizations lamenting their inability to find the right talent, and, on the other side of the problem, too many talented people complaining about meaningless or uninspiring jobs.
It seems that, with the exception of professional sports, the talent market is far from efficient, highlighting a pathological mismatch between supply and demand of talent. The equivalent in the realm of relationships would be the majority of people being either single or unfulfilled by their romantic partners – which is not too far from reality.
Inefficient markets are always ripe for disruptions, and innovation with impact always bridges the gap between supply and demand. In order to achieve this, it may be necessary to redefine, or even reimagine talent. Surprisingly, our notions of talent have not evolved to keep up with the times. When university credentials have become disconnected from job-relevant knowledge, hard skills quickly become outdated, and what we know is less important than what we can learn, organizations are often left looking for talent in all the wrong places. This also harms their desire to create a diverse and inclusive workforce. When your main talent currency is still the resume, and the value of a resume depends on outdated talent currencies like college qualifications or past experience, it is hard to avoid hiring the same type of people over and over again, optimizing for “culture fit” rather than diversity.
In this context, social media emerges as a promising alternative to the dominant currency for talent. Its data acts as a talent bitcoin capable of redefining human capital more inclusively and meritocratic. Our social media activity already reveals a great deal of information about our deep character traits, precisely the type of stuff employers need to know (and at times also want to know) before they decide to hire us.
Here are just some examples derived from the hundreds of independent scientific studies in this area.
- Facebook: As the unfortunate Cambridge Analytica fiasco showed, our Facebook data predicts not only our consumer preferences, but also our personality traits, which have long been associated with different vocational preferences, and work-related competencies. Through publicly available data on people’s conscious self-presentation (e.g., groups, media content, and posts they like) algorithms can estimate their intelligence, curiosity, and extraversion, which have been consistently linked to job performance and underpin many definitions of talent. In fact, talent is mostly personality in the right place.
- Twitter: Unstructured data, such as text, has been consistently linked to deep character dimensions, and Natural Language Processing has shown to be a reliable form of AI to translate people’s communicational style into a predictive model of their personality. The words you use and how often you use them can tell us whether you are calm or emotional, driven or laid back, trusting or skeptical, positive or negative, etc. Incidentally, these dimensions of personality don’t just predict talent for common jobs, but also leadership.
- Instagram, YouTube, and TikTok: The videos people upload on social media convey millions of signals about their verbal and non-verbal communication, including physical properties of their voice, and facial expressions that have been associated with psychological traits relating to their emotional intelligence, empathy, and sociability. To be sure, the algorithms are far from perfect, but with enough data, they are capable of outperforming humans on emotion perception, especially when analyzing strangers.
- Spotify and other smartphone and sensing data: Music consumption, overall phone activity, app usage, and day vs night activity also reveal core ingredients of talent, such as whether you are confident, optimistic, risk-averse, and hard working. Consider this study which mined 17.6 million Spotify songs to correctly infer listeners’ personalities, moods, and behaviors. Although these signals are not yet used to determine people’s fit for a job, you could picture a future in which more people pretend to listen to classical music or jazz (and delete Justin Bieber from their playlist) in order to improve boost their personal talent bitcoin.
There are obvious ethical concerns (and sometimes legal constraints) that limit the use of social media in hiring and talent identification, which have fortunately stopped organizations from using AI to scrape candidates’ feeds to infer their talent or potential.
However, it could be addressed. All it would take is to adequately brief candidates, invite them to engage in an informed and transparent transaction where they decide to opt-in. That would allow algorithms to translate their data into a talent bitcoin, and they are free to decide whether to keep this information private and trade it as a currency with potential employers and recruiters. This might increase the motivation to fake it and manage impressions on social media, but it’s not as if people are currently being themselves online. We already self-edit to curate our digital reputation in order to please or impress others.
Besides, our social data has already been commoditized to show us more relevant ads that make us want to buy things we didn’t need. Surely a meaningful job would be at least as useful?
Perhaps more importantly, it would be useful to acknowledge that unethical uses of social media for talent are quite common today, without any intervention of AI. The first is that hiring managers and recruiters are already looking at these data, as they will often Google a candidate to inspect not just the skills and work experience they report on LinkedIn, but also the pictures they share on Instagram, the groups they like on Facebook, and the things they say on Twitter.
This information is obviously public, but it is unlikely that candidates are fully aware of the signals they send, and the potential career consequences of making this information public. It is also unlikely that recruiters are able to consistently infer candidates’ talent, paying attention to the right signals, and ignoring the information they should ignore (e.g., gender, age, attractiveness, and race). First impressions are heavily influenced by stereotypes and prejudices, and no amount of unconscious bias training can make human recruiters forget that a candidate is female, old, of a different race or ethnicity, or unattractive.
Since current hiring practices are far from objective, and typical evaluations of talent are neither unbiased nor tremendously accurate, it wouldn’t take much for algorithms and social media to rival traditional measures of talent in terms of both trust and perceived value. Just as one day Bitcoin may be trusted more than national currencies.