Why We Need a Conversation Around Open-Source AI

Recently, Mark Zuckerberg (Meta& Facebook Founder/CEO) and Daniel Ek (Spotify Founder/CEO) co-authored an article on embracing open-source AI (posted here).  At first glance, these kings of monetizing digital content seem like strange proponents of free, publicly accessible tech.   But this short read confirms that these CEOS are also innovation leaders.  Their points seem to grasp that, as AI is quickly transforming the world around us, there is a real need to proactively lean into the all the cheers and fears that come along with the perception & reality of the seismic shift in how machines & humans interact. Change is coming…let’s sort it out together.

There is not much debate that Artificial Intelligence (AI) is no longer just a buzzword—it's something that is becoming part of our daily lives. But as AI continues to grow and evolve, the conversation is more that how it should be developed & pragmatically used but also controlled & governed.  Though AI is in its relatively early stages, the headlines are already filled with stories about AI’s negative impact:  Job Displacement, Bias & Discrimination, Privacy Invasion, Misinformation & “Deepfakes” and Security Risks.   Are we adding to the problem by creating open access AI for anyone to see, change or distribute?   Considering the history of open-source technology, I’d like to submit a few points in favor of open-source AI.

Why Open-Source?

Spark Innovation

Innovation is often the lifeblood of any tech field, and AI is no exception. Linux, Apache, and TensorFlow are just a few examples where community-driven development has led to powerful and widely-used technologies. Developers and researchers around the world build on each other's work rather than starting from scratch, speeding up innovation.   For example, when Google made TensorFlow open-source technology, it quickly became a favorite among AI developers and sparked a wave of research and new applications. With an open platform, developers could tweak it, add new features, and fix bugs, creating a snowball effect that accelerated progress in AI.

To cut to the point, open-source technology encourages experimentation, trying out new ideas freely that lead to breakthroughs that might not have happened in a more restrictive environment.

Build Trust with Transparency

AI can sometimes feel like a "black box”, not having insight into how these systems make decisions. How can we trust something we don’t understand?

This is where open-source AI can make a big difference. When the code behind AI is open, anyone can look inside to see how it works. This transparency is crucial for building trust. Whether it's independent researchers, regulators, or just curious members of the public, having the ability to inspect AI systems helps ensure they’re working as they should be, and aren’t biased or discriminatory.

For example, there have been publicized instances where proprietary AI models were found to be biased against certain racial or gender groups. With open-source AI, more eyes can spot these biases early on, and the community can work together to address them, making the technology fairer for everyone.

Make AI More Secure

Security is always a top concern when it comes to technology, and AI is no different. Proprietary AI models can be a bit of a mystery when it comes to security—since the code is closed off, it’s hard to know where the vulnerabilities might be.

Though it seems like a security oxymoron, Open-source AI can be more secure as it benefits from the "many eyes" principle.  With more people able to look at the code, it’s more likely that someone will spot potential security issues before they can be exploited. This is similar to what we’ve seen in other open-source software, where vulnerabilities are often identified and fixed faster than in closed systems.

Additionally, open-source AI can help prevent the risks that come with having a few big companies control the technology. When AI is proprietary, it puts a lot of power in the hands of those who own it, which can lead to monopolies and increased security risks. Open-source AI spreads that power around, allowing smaller companies and individual developers to contribute, which in turn makes the whole ecosystem more secure.

Promote Inclusivity

AI should be for everyone, but that’s not always the case when the technology is proprietary. High costs and restricted access can create barriers, especially for people in developing countries, small businesses, or underrepresented communities.

Open-source AI helps break down these barriers. When AI tools and models are open-source, they’re available to everyone, regardless of where they live or how much money they have. This democratization of AI technology means that a wider variety of people can contribute to its development and use, leading to solutions that better reflect the needs of all kinds of people.

For example, in healthcare, open-source AI could enable doctors and researchers in low-resource areas to create solutions tailored to their specific needs. Instead of being stuck with one-size-fits-all tools, they can modify open-source models to fit their local context, whether that means adjusting for different languages, cultural practices, or prevalent health issues. The same is true for education, where open-source AI can help bring advanced learning tools to students and teachers everywhere, not just in the most developed parts of the world.

Support Ethical Development

As AI starts to play a bigger role in our lives, we need to be more diligent about how it’s developed and used. There are big ethical questions around AI—how to prevent bias, protect privacy, ensure accountability, and avoid misuse. Open-source AI can help address many of these concerns:

·        Simplifies developers accountable. When AI systems are transparent, it’s much harder to hide unethical practices. The open-source community can spot potential ethical issues and push for changes to make sure AI technologies are developed and deployed responsibly.

 ·        Easier to align development with shared ethical values. When everyone can contribute, it’s more likely that the AI model will reflect a broader consensus on what’s right and fair. For example, the open-source community has played a big role in developing privacy-preserving technologies like federated learning, which allows AI models to learn from data without actually collecting that data in a central location. This is crucial for protecting privacy while still enabling AI to get better.

 ·        Helps prevent the misuse of AI. Proprietary AI systems can sometimes be used for harmful purposes, like surveillance or spreading disinformation. With open-source AI, communities can develop tools that counter these risks, such as detecting deepfakes or identifying bias in AI systems.

Empower Education and Skill Development

As AI becomes more important in the job market, the demand for AI-related skills is skyrocketing. But not everyone has access to the resources they need to learn these skills, especially in places where educational opportunities are limited or expensive. Open-source AI can help bridge this gap.

Open-source AI platforms and tools are incredible resources for anyone who wants to learn. Platforms like GitHub, Kaggle, and Google Colab give learners access to open-source projects, datasets, and tutorials, allowing them to gain hands-on experience in AI development. By working on real-world projects, learners can build practical skills and create portfolios that can help them in their careers.

Open-source AI also promotes a culture of continuous learning and knowledge-sharing. In the open-source community, people are constantly exchanging ideas, best practices, and new techniques, which keeps the field dynamic and helps everyone stay on the cutting edge.

And it’s not just for self-learners. Universities and schools can use open-source AI tools to enhance their curricula, providing students with hands-on experience that’s often more valuable than traditional classroom learning. Open-source resources can also help create more affordable and accessible educational materials, making it easier for people everywhere to learn about AI.

Set Industry Standards and Ensure Interoperability

For AI to really take off across different sectors, we need common standards and ways for different systems to work together—otherwise known as interoperability. Open-source AI can be a driving force in setting these standards (think about Linux and Apache Web Server as examples).   Beyond technology compatible, common standards will be crucial across industry types (such as  healthcare, where AI-driven tools need to work seamlessly with things like electronic health records to be truly effective).

Open-source AI also helps prevent the fragmentation of the AI ecosystem. When different AI systems are incompatible with one another, it creates silos and inefficiencies. By encouraging the development of interoperable solutions, open-source AI helps create a more cohesive ecosystem where different tools and technologies can work together more effectively.

Encourage Cross-Sector Collaboration

AI isn’t just about one industry or one use case—it has the potential to impact almost every part of our lives, from healthcare and finance to education and the environment. Addressing the complex challenges that come with AI requires collaboration across different sectors, and open-source AI provides the perfect platform for this.

When AI technologies are open-source, it’s easier for experts from different fields to work together on solutions. For example, in healthcare, doctors, data scientists, and AI developers can collaborate on open-source projects to develop tools that improve patient care. In environmental science, open-source AI can help conservationists, climate scientists, and technologists create tools for monitoring and mitigating climate change.

Open-source AI also makes it easier for people outside of the tech world to get involved. Social scientists, ethicists, and legal experts can contribute to open-source AI projects, helping to address the social, ethical, and legal implications of AI. This kind of cross-disciplinary collaboration is essential for ensuring that AI is developed in a way that benefits everyone, not just a select few.

Lower Costs and Break Down Barriers

AI can be expensive, and when it’s locked behind proprietary walls, the costs can create significant barriers to entry. This is especially true for startups, small businesses, and individuals who may not have the financial resources to access proprietary AI tools and services. Open-source AI helps to level the playing field.  By making AI tools and models freely available, open-source AI allows developers and organizations to build and deploy AI solutions without expensive licenses or subscriptions.

Open-source AI also offers more flexibility. Proprietary AI systems are often rigid and hard to customize, but open-source AI can be adapted to fit specific needs without the constraints of a commercial license. This flexibility is crucial for developing AI solutions that are effective in a variety of contexts, from local businesses to global initiatives.

Wrapping It All Up

Open-source AI isn’t just about making AI free.  Though it seems contrary, the open-source model has a history of making technology better, fairer, more secure and more accessible to everyone.   As there is no consensus yet among technology professionals, it not surprising that the discussion is even more heated among the general population that has been thrusted into this new paradigm. Ready or not, AI is here and will continue to shape our future. Embracing the open-source model could be the way to ensure that the benefits of AI are shared widely, responsibly and ethically.   It’s why we need a broader conversation around open-source AI.