Let’s Hear It From The Experts

7: Let’s Hear it from the Experts

In the rapidly evolving world of AI, experts are speaking up about the power and potential of crowdsourcing—and the ethical concerns that come with it. Here’s what some of the leading voices have to say.

7.1 Cassie Kozyrkov (Google AI Lead): Why Diversity Matters in AI

“Crowdsourcing democratizes AI training, ensuring broader representation and inclusivity.”

Cassie Kozyrkov is a strong advocate for making AI accessible to everyone. She believes that while AI systems are advancing at lightning speed—powered by immense memory capacities and computational scales—we need to ask: Who’s in control? And who truly benefits from AI’s progress?

Cassie emphasizes that building complex AI systems shouldn’t be reserved for those with specialized degrees. Instead, anyone, regardless of technical expertise, should have the chance to contribute to AI development.

“Diversity isn’t just a nice-to-have; it’s a must-have,” she states. For AI to work effectively across a broad range of users, it needs contributions from people with different life experiences and perspectives. By welcoming diverse voices, crowdsourcing makes AI smarter, fairer, and more inclusive.

7.2 Jeremy Howard (Fast.ai): The Community-Driven Approach

“Community-driven AI models outperform corporate-trained ones in terms of adaptability and fairness.”

Jeremy Howard is on a mission to make AI accessible to everyone—not just big corporations. As the co-founder of FastAI and Answer.AI, his goal is to create tools that benefit society and empower everyday people to harness the power of AI.

By building collaborative platforms that welcome learners and innovators, Jeremy envisions a future where AI extends beyond simple chatbots and coding assistants. From healthcare solutions to creative problem-solving, his approach focuses on expanding AI’s potential to serve real-world needs.

“The journey of AI is just beginning, and community contributions are what will drive it forward,” Jeremy explains. His dedication to inclusivity and accessible AI tools continues to shape how people interact with and benefit from AI.

7.3 Andrew Ng (DeepLearning.AI): Open-Source Data Is the Future

“The next wave of AI innovation will come from diverse, open-source data contributions.”

Andrew Ng believes that AI’s future depends on democratizing access to quality data. Unlike the traditional model-centric approach, Andrew emphasizes a data-centric philosophy where well-organized, high-quality datasets are prioritized over sheer volume.

Crowdsourcing, he says, offers a practical way to enhance AI’s capabilities by gathering real-world data from diverse sources. “If we want AI to truly benefit everyone, then everyone should be part of its development,” Andrew asserts.

He identifies two key barriers to AI’s widespread adoption: small datasets and customization difficulties. Crowdsourcing can overcome these challenges by pooling knowledge from various backgrounds to build more accurate, inclusive, and powerful AI systems.

7.4 Timnit Gebru (AI Ethics Researcher): Fairness Before Innovation

“AI crowdsourcing risks perpetuating exploitation unless fair compensation models emerge.”

A relentless advocate for ethical AI, Timnit Gebru has been vocal about the industry’s shortcomings when it comes to fairness and transparency. Her concerns are rooted in the low pay and mental strain that many data annotators face—issues she believes are often overlooked by corporations eager to scale their AI models.

“Talking about ethics is not enough; we need institutional and structural changes,” she insists. The exploitation of low-wage data contributors is a pressing concern, and she calls for establishing fair compensation models that recognize the value of their work.

Timnit also highlights the importance of sourcing data from diverse contributors to minimize biases and improve AI’s ethical integrity. To her, crowdsourcing’s potential can only be realized when ethical principles are prioritized over profit.

7.5 Gary Marcus (AI Critic): The Dangers of Crowdsourced AI Data

“Publicly sourced AI datasets often lack the depth and rigor needed for critical applications like medicine.”

Gary Marcus, a cognitive scientist and seasoned AI critic, believes the industry’s rush to embrace crowdsourcing comes with significant risks. His main concern? Many AI systems rely on shallow, publicly sourced datasets that are ill-suited for high-stakes applications such as healthcare and financial services.

According to Marcus, today’s AI models are often unable to distinguish between crucial and trivial tasks. The core problem, he argues, isn’t just deep learning itself but the methods used to train AI systems.

“The lack of rigor and transparency should trouble everyone,” he asserts. For AI to truly be effective, he insists that crowdsourced datasets must be curated with far greater attention to detail and reliability.

These influential voices offer a powerful glimpse into the future of AI and crowdsourcing. From increasing inclusivity and improving fairness to confronting ethical dilemmas head-on, experts are calling for a more transparent and collaborative approach. As AI continues to evolve, balancing innovation with ethical practices will be essential to creating systems that benefit everyone.

Contributor:

Nishkam Batta

Editor-in-Chief – HonestAI Magazine
AI consultant – GrayCyan AI Solutions

Nish specializes in helping mid-size American and Canadian companies assess AI gaps and build AI strategies to help accelerate AI adoption. He also helps developing custom AI solutions and models at GrayCyan. Nish runs a program for founders to validate their App ideas and go from concept to buzz-worthy launches with traction, reach, and ROI.