AI and HR: The Disrupted Enterprise – Emmanuelle BLONS

Artificial Intelligence: Nothing More Human

Emmanuelle began by counter-intuitively redefining artificial intelligence: “There is nothing more human than artificial intelligence.” AI relies on three elements: a lot of data (collected and taught by humans), an algorithm (coded by a developer), and machine learning (often supervised by a human). This perspective helps us understand that everything good, but also everything bad, can go into AI.

AI intervenes in HR in two main aspects: how HR can use AI to improve its processes (recruitment, performance management, training), and how HR can prepare the company for the arrival of AI in other processes, particularly in terms of change management and personal data protection.

Predicting and Preventing Attrition with AI

A concrete example illustrates AI’s potential for human resource management: the prediction of potential departures. By analyzing criteria such as tenure in the position, recent promotions, interactions with the manager, and LinkedIn activity, AI can identify employees at risk of leaving the company.

However, the machine should never have the final say. The manager must exercise critical thinking and interpret the results. For example, increased activity on LinkedIn can signify a job search, but also active participation in a company ambassador program. Human intelligence must always be exercised to challenge and interpret AI predictions.

Optimizing Recruitment with AI

AI can also revolutionize recruitment by seeking out completely atypical profiles that traditional recruiters would not have identified. L’Oréal’s example is telling: AI suggested recruiting lawyers for sales positions, and these profiles were extremely successful because AI had identified transferable skills that a human recruiter would not have had time to delve into.

AI can also analyze facial micro-expressions, lexical fields, and voice tone during video interviews, providing additional information to the recruiter. But once again, these tools must complement, not replace, human judgment.

Engaging Teams with AI

AI can also contribute to employee engagement in two ways: sentiment analysis to understand employees’ state of mind from surveys and internal forum discussions, and empowering employees in their careers by offering them training and positions suited to their skills, as Natixis does with its matching tool.

This approach allows employees to interact directly with AI, define their skill levels, and receive personalized suggestions, thus creating an experience similar to Netflix, where the algorithm learns user preferences.

Risks and Biases to Avoid

Emmanuelle warned against three levels of risk: biased data (like Amazon’s example of recruiting only men), biased algorithms (coded by undiverse teams), and biased machine learning (trained by humans with unconscious biases).

To avoid these pitfalls, two pillars are essential: transparency (being able to “lift the hood” to understand the algorithm) and diversity (having diverse teams to code, test, and interpret results). Ethics by design must be integrated from the design phase.

In a nutshell

Here are the key takeaways from this Masterclass with Emmanuelle Blons:

• AI as a Complement, Not a Substitute
  Use artificial intelligence to optimize decision-making and free up time, while retaining human judgment for strategic decisions.

• Preventing Attrition and Optimizing Recruitment
  AI can identify at-risk employees and discover atypical profiles, but the manager must always exercise critical thinking.

• Transparency and Diversity to Avoid Bias
  Ensure that AI development teams are diverse, and be able to understand and challenge the algorithms used.

Don’t miss the next Masterclass!

• Prospecting: The Revolutions – Toinon Georget
• Growth Financing – Julien Zerbib
• LinkedOut – Mobilizing for the Inclusion of the Most Disadvantaged – Jean Marc Potdevin