Exploring the Synergies Between Open Source & AI Adoption
An Open Source Perspective on the Era of AI
Collectively, we’re fascinated by AI’s power and curious of its potential. At the same time, it gives us pause — for many valid reasons. When leveraged thoughtfully and responsibly, it can deliver so much opportunity. Enterprises are at a crossroads: Many are frozen by AI paralysis but, to compete at the velocity that AI is evolving, they need to start adapting to this strange new world, now.
What businesses are facing today with AI adoption has similarities to what organizations faced with open source years ago. There’s a lot we can learn and there’s significant overlap to explore in how AI and open source can be used together. As such, we’ve compiled an open source perspective on the era of AI to consider as you navigate the path ahead…
Open Source & AI’s Evolution into Commercialization
If you consider the last 20 years, AI and open source have developed and matured along similar paths. Open source — and the way in which the methodology works — relies on collaboration at scale, bringing experts together to share best practices and build better solutions and projects. Similarly, AI was born out of research and development activities, with academic institutions and companies investing significant time and money around the world.
Both AI and open source endeavors reached a point where their creations morphed into an attractive offering for the commercial world. Looking at the two models, these now enterprise-grade products and solutions emerged out of decades worth of research and innovation. ChatGPT is a great example. OpenAI was set up to be entirely open as an AI research lab and, following significant investment, shifted to align to a proprietary entity. Elastic Search is another example of a technology platform that started as an open source solution and in recent years, has shifted from permissive to more restrictive licensing conditions.
How Open Source Can Enable Ongoing AI Development
Some institutions are exploring open source as an enabler of ongoing AI development. Open source leader the Linux Foundation, for example, recently formed the LF AI & Data Foundation recognizing promise in aligning open source and AI through methodologies centered around transparency and accountability that ultimately foster the development of trustworthy AI systems. The foundation’s existing approaches, methodologies and ways of working enable them to effectively build solutions and governance for these systems. The Linux Foundation’s overarching goal with this open source AI and data ecosystem is to lower the barrier of entry to AI development by bringing together AI projects that have potential and applying their frameworks to progress toward the next stage of commercialization.
Embracing Adoption & Overcoming Regulatory Barriers
Highly regulated industries are the first to come to mind among those that are the most cautious about AI adoption. When you think about how AI relies on massive data banks to train, refine and enhance its models, this makes many businesses reluctant to adopt AI because it potentially means putting sensitive data about their operations into the public domain.
For example, banking institutions may be averse to disclosing how they assess risk or share details related to their investment strategies. As open source grew in popularity, these same banks were resistant to adoption as well. However, as open source has matured and financial services companies have seen the value in open collaboration, the growth of organizations, such as the Fintech Open Source Foundation (FINOS), have helped champion the value that open source provides. The use cases that AI offers to the financial services industry include topics such as intelligent fraud detection, automated investing and heighted customer engagement through chatbots and sentiment analysis.
We see common trends in healthcare, too. There are many non-profits and organizations focused on improving the world and the lives of people through the power of data, leveraging computer imaging and data modelling in cancer research for example. Having the ability to populate complex medical data into open AI tooling is worth exploring as these organizations no longer must rely on investing and building their own infrastructure and instead can take advantage of existing offerings. AI has the possibility to drive rapid positive changes that can enhance the quality of life through more accurate diagnoses, including opportunities to map and track the spread of infectious diseases to help prepare for future pandemics.
In the automotive industry, open source plays a big part in building the software-enabled vehicle. An increasing number of automotive companies are considering how AI and machine learning will become a core capability in self-driving, autonomous vehicles, integrating with cities of the future and manufacturing automation including drones and robotics.
Finally, thinking from a government perspective, there's even more hesitancy around policy, legislation, governance, security and data privacy — to name a few — when considering both open source and AI. So, how do organizations and enterprises become the enablers and change agents to help inform and shape policy? The White House, for example, has been inviting industry experts to share their perspectives around open source and software bill of materials to help shape national policy. Additionally, global data-orientated businesses are being tasked to share details on how they’re leveraging and using data.
Some of these obstacles make it difficult to envision where to even begin with AI adoption, but there's certainly an opportunity for enterprises to be advocates for data transparency as they enter the conversation around AI usage.
The Path Ahead
Both AI and open source are deeply rooted in a community of contributors — from academic institutions to companies to individuals — helping to accelerate research and development. Both technologies would never have reached the level of maturity they did, nor in the timeline they did, without this community support and engagement. By leveraging the ways of working and frameworks used for open source, along with open source solutions to enable AI development, we're seeing significant advances particularly over the last 18 months.
Moving forward, tapping into all that AI can offer will fortunately not require reinventing the wheel for this maturing technology. Sharing data through global repositories and attracting ongoing contributions from leaders in this domain will provide positive results that accelerate adoption. Ongoing investment and increased attention around the formation of industry bodies (such as those mentioned earlier) will continue to bring together academics, business leaders and those simply passionate about AI to drive synergies, share ideas and establish new approaches to use and enhance AI.
This is the easiest way for enterprises to overcome the latest AI disruption and begin to build applications, accelerators and data sets that will redefine the value delivered to customers while improving the attractiveness of vendor offerings. By incorporating AI into our ways of working, companies can transform the entire enterprise experience.
