Artificial Intelligence's Intelligence Level Equals the Quality of Information It's Given.
I recently discussed at Web Summit and AWS' CHRO Summit about a topic that is crucial in understanding the advancement of businesses in the era of artificial intelligence (AI): the significance of human data in fostering human intelligence. While AI has the potential to revolutionize and transform, its capabilities are fundamentally limited by the quality of data we provide. According to McKinsey, biases in data, whether due to unbalanced collection practices or human judgment, can perpetuate and even intensify societal disparities when AI systems analyze, learn from, and act upon such data, leading to far-reaching and unintended consequences.
In essence, feeding AI faulty, insufficient, or superficial data restricts its potential and rather amplifies our own biases and mistakes.
AI: A Powerful Magnifying Glass (If It Gets the Right Data)
Humans have accomplished remarkable feats even with the most basic tools. The Apollo 11 moon landing was achieved with a computer less powerful than modern toaster ovens. Computers, as Steve Jobs once put it, are like "bicycles for the mind" because they amplify our abilities. By this logic, AI is a powerful magnifying glass for the mind, enabling it to process data and expose insights that humans could never perceive on their own.
However, AI only learns what it is taught. If its foundation is flawed, AI will reflect those flaws right back at us. This is why 80% of AI projects fail, according to a study by the Rand Corporation.
Even AI recognizes its limitations. For instance, when asked how it could assist in writing job descriptions, ChatGPT outlined its abilities but added this caveat:
"AI-generated descriptions might require human oversight to ensure that the specificities of the job and company culture are accurately represented. In certain cases, AI might overlook the emotional or aspirational factors that are important in attracting top talent."
The issue isn't AI itself—it's the data we're feeding it. Traditional metrics like turnover rates and engagement scores provide a snapshot of employee behavior, but they lack explanation. Performance reviews focus on skills and objectives but fail to comprehend the actual process of work. This discrepancy leads to faulty decision-making, perpetuating bias and limiting AI's capacity to deliver meaningful insights.
On the other hand, human data or the real-time, nuanced interactions and feedback between individuals offers a more genuine, authentic understanding of workplace dynamics. This is where recognition data makes a significant difference.
Recognition: The Cornerstone of Human Data
Recognition data refers to the insights gained from analyzing messages of appreciation and gratitude exchanged between employees. It is one of the purest forms of human data because when someone expresses gratitude or praise, it reflects not only the work they do but also the value and impact of their efforts.
Consider this real recognition note:
"John, I wanted to express my appreciation for the thoroughness and precision of your reports during client presentations. Your collaboration with the sales team is a remarkable example of cross-functional teamwork."
The right AI setup can associate qualitative performance indicators in written feedback—such as thoroughness, precision, collaboration, and teamwork—with John, and compare his performance to that of his colleagues. It can identify John as a potential mentor or partner for new hires and recognize which qualities contribute to the best business results.
Collectively, these recognition moments—where employees acknowledge and celebrate each other's contributions, achievements, and progress—produce a wealth of insights into performance, collaboration, and culture. This human data provides a comprehensive understanding of how individuals interact, lead, and engage within an organization, offering insights into areas that traditional HR tools often struggle to measure.
Utilizing Human Data
Now imagine aggregating this data across numerous organizations, resulting in billions of data points about how work actually occurs. To put this into perspective, a million seconds is approximately 12 days, while a billion seconds span 32 years—nearly an entire career's length. The outcome is a focused and extensive collection of human interactions that can be analyzed by AI tools to reveal best practices, discover hidden gems or quiet influencers, and even quantify the return on HR investments.
Suppose you ask AI to identify the most influential person in a recent product launch. Traditional data may suggest a project manager or designer, but AI trained on recognition data might surface an unexpected result: the office manager, Emily. Emily was not involved in the meetings or the project plan, but her behind-the-scenes support—scheduling, approvals, and organization—kept the project on track.
Recognition moments capture these unseen contributions. They function like the memory orbs from Pixar's Inside Out—not just storing emotions but revealing crucial insights about who contributes, how they perform, and what behaviors drive success.
I'll give another example: In a recent experiment, I asked AI to identify the greatest flight risk at Workhuman. The result? Me. As a CEO, I do not often receive recognition, so the data flagged me as disengaged. While I am not planning to leave, this situation underscores how AI can expose unexpected insights, inspiring proactive actions.
The forward-thinking organizations of tomorrow will leverage human intellect to enhance their human-derived data, transforming daily encounters into valuable actions that fuel culture, invention, and success. This transition markedly shifts us from gut feelings to intelligence, from presumptions to plans. Recognition data functions as the propellant powering this voyage, empowering HR leaders to assume their proper roles as strategic advisors in the executive suite.
In my opinion, genuine greatness stemms not from machines but from individuals. This means that human intellect is crucial in unlocking the highest potential within an organization. Embracing AI extends the capabilities of human intellect should be our approach to the evolving job landscape. Therefore, let's zero in on the data that holds significant value—human data.
- To maximize the potential of AI in fostering human intelligence, it's essential to avoid feeding it faulty, insufficient, or superficial data, as this will only reflect and amplify our own biases and mistakes.
- In the discussion about AI and human data, it's important to remember that while AI is a powerful tool for analyzing and interpreting data, its capabilities are fundamentally limited by the quality of data we provide.
- The effectiveness of AI in recognizing job-related qualities and potential mentors or partners relies on the data it is fed. Traditional metrics may provide a snapshot, but human data from recognition moments offers a more holistic understanding of workplace dynamics.
- Amazon Web Services (AWS) CHRO Summit and Web Summit provided a platform to discuss the significance of human intelligence in the era of AI, emphasizing the need to focus on high-value human data to unlock the full potential of AI and augment human intellect.
