Now OpenAI is coming for LinkedIn

As press releases go, it was a mic drop moment.

On September 4, OpenAI announced that it was coming after LinkedIn. And it was coming after job search specifically.

For those unfamiliar with the story, it was reported across multiple sources that not content with torpedoing Google, education, interpersonal relationships, and entire industries such as psychiatry or management consultancy (both smugly considered untouchable until about 15 minutes ago), OpenAI is now coming after the age-old industry of human resources.

Specifically, the $300 billion company wants to streamline the process of hiring by automatically matching candidates to roles via context alone, without the need for titles, résumés, applications, or screening—a fully automated, frictionless match between resource and need, run at zero cost in a fraction of the time it would take Becky from HR to book meeting room 14B next Tuesday, but with near-infinite scope and efficiency.

It’s yet another cold, logical extension of artificial intelligence’s purported ability to reduce human intuition down to a series of predictable algorithmic factors.

LinkedIn Dethroned?

Now, let’s be fair. OpenAI’s proposition assumes (probably rightly) that virtually all future roles will require some degree of AI knowledge or fluency. Moreover, the press release breathlessly proposes an OpenAI Academy, one that will train the machine-picked workers of the future on how to best serve their artificial overlords.

But will this really be enough to dethrone LinkedIn? An incumbent platform of over 1 billion people who now spend the bulk of their time not in job search or hiring, but in perpetual AI comment-discussions about the nature of “authenticity” while posting Canva-selfies of their weekend getaway to Puglia?

Perhaps we should first recognize that human resources is one of the oldest and most safeguarded functions within the corporate world, holding out defiantly as personal assistance, finance, and even marketing have been variously usurped by automation.

It’s a curiously old-fashioned industry in many respects, relying on judgments about physical appearance and behavioral norms such as eye contact and punctuality—characteristics that in any other sector would be subordinated to one’s ability to deliver value or output.

But recently, HR has gone through something of a transformation, and it’s safe to say that LinkedIn has been at the heart of that.

In fact, the world’s largest business platform been an unwitting catalyst for an onslaught of AI-driven fraud—both from candidates (either building fake AI profiles or even replacing themselves in Zoom interviews with omniscient AI avatars), or from dubious recruiters (leveraging advanced AI tools to build fictitious companies in order to solicit payment or personal details).

The market is now flooded with AI cover letters, AI résumés and AI-curated portfolios, while on the other side of the table, the screening process is often (and increasingly) handed off to AI scanning software, in order to save valuable time filtering out unsuitable candidates before proceeding to a final interview—one that may ultimately be conducted and processed by an AI avatar.

Surely then it’s simply a short step to full automation, one that removes these very ghosts in the hiring machine?

Perhaps. But I’d suggest that any scholar of organizational psychology would hold a very different view.

The intangibles

We can look, for example, at evidence highlighting the predictive power of nonverbal cues such as vocal tone, facial expressions, and body language in forecasting later job performance. A 2022 Harvard Business Review piece, on the other hand, cautioned that automated interviews may be prone to missing these types of “intangible” characteristics—potentially valuable employee attributes (say, curiosity or resilience) that might only be registered via live human interaction.

Yet suffice to say, these same factors are also deeply subjective. Economists have long shown that reliance on intuition and first impressions can also be a major source of bias in recruitment, perpetuating disparities around race, gender, and class.

AI proxies

Whatever the case, the stark reality is that we now find ourselves in an era where a growing majority of both candidates and recruiters are using AI as proxies. And in this world, the very notion of authentic observation—whether helpful or not—is starting to erode.

But—I hear you cry as one—surely HR isn’t just about hiring? It’s about people management, professional development, organizational health, and so many other things?

Yes—except that of all these functions, hiring has historically required the highest amount of dedicated resource, training, and physical presence, far outstripping what can now be done by AI in areas such as coaching, psychometric testing, and even therapy. If we delegate the very function of organizational building itself, what remains must necessarily become diminished, at least in part.

To circle back then to OpenAI, perhaps iterative machine learning will finally win out. Perhaps time might tell us that the match between employee and role is near-flawless in an AI, dehumanized world. And yet, it may also be this very unpredictability of human nature that will represent the biggest loss to a future workforce.

And as to the suggestion that OpenAI’s hiring platform will take down LinkedIn itself?

If it can somehow balance hiring efficiency with the high level of fluffy, eye-pleasing pseudo-business interactions that now so predominate on LinkedIn itself, perhaps.

But to do that, you’d need to first understand a little of the idiosyncrasies of human nature, in order to build (and foster) an appropriately “sticky” forum—if only for the reason that “business” is a gateway into so many other areas of recreational and emotional life.

And doing that, I’m afraid to say, still requires a deeply human sensibility.

View the full article