The future of AI is Open

Without question, AI is primed to change the nature of how we work, with the potential to help businesses unlock $16 trillion in value by 2030. This means enterprises will need to reckon with how best to leverage this tool in the future. However, like other revolutionary technologies, the printing press and the internet among them, there’s plenty of potential for misuse as well.

While previous redefining technologies like the printing press, the internet and social media, have advanced the way we work, play, learn and connect, these technologies have also needed updated policies when threatened by improper practices. In this respect, AI is no different. The flip side to knowledge sharing, education, journalism and the general spread of novel ideas means AI may be vulnerable to disinformation, bias and inaccuracies.

While poised to revolutionize all facets of society, artificial intelligence is rife with both cautious optimism and skepticism from consumers and businesses leaders alike. Consumers often don’t realize how much they are already using AI in their day to day lives. A 2023 PEW Research survey which found that 30% of Americans, when asked to recognize up to six forms of AI, including wearable tech and product recommendations, couldn’t do so. So how do we bridge the gap between AI understanding and AI adoption? As this technology evolves, acceptance may seem a prudent first step. A step that also requires collaboration.

“So if you look at the ecosystem of AI, it transcends from technology companies, software and hardware, to application developers to academia, to startups, to national labs across the globe. This is the ecosystem that we mean, that we need AI to be open” says VP of IBM Research AI, Sriram Raghavan. This sentiment is at the heart of the AI Alliance, an international community of more than 100 organizations bringing together top AI scientists, researchers, policy experts, business leaders, and innovators focused on accelerating and disseminating open innovation across the AI technology landscape.

Like any technology, AI is a tool. It can help individuals and enterprises alike to create value by doing away with labor intensive tasks. Raghavan elaborates on this and says, “at its core, AI as a technology is about two things, getting insights out of data, and automation.” By offloading laborious tasks, businesses can create value for their customers in other ways, like developing new products or offering a more personal approach with clients. It’s about making sense of data and doing automation or improving productivity. By making sense of that data, businesses hope to convert that comprehension into actual dollars.

“An open ecosystem has an opportunity to influence the way this technology evolves,” says Raghavan. “All of the stakeholders in this ecosystem are able to participate in the creation, adoption, benchmarking, and evaluation of models. The development of AI technologies should not be determined only by the few. Whenever you have a profound technology, it’s critical that its use and development is unconstrained in the open.”

With the acceleration of innovation, open models will be a defining force in the evolution of AI technology now and in the future. That’s why in collaboration with Red Hat, IBM recently launched InstructLab - a new and unique open-source project which places large language model (LLM) development into the hands of the open-source developer community. Collectively, that community can contribute new skills, knowledge and information to any LLM. Going further, IBM open-sourced a family of its most capable Granite models which help transform and optimize applications for enterprises. The goal with open-sourcing these models is to make tasks such as coding as easy and accessible as possible, for as many developers as possible.

The implications of AI are many, from enterprises to the individual or end user of the technology. Taking this into consideration, if the printing press or the internet had only been in the hands of a few companies or institutions, the world as we know it would look very different. Burgeoning technology like AI is crucial for businesses, and people alike, to acknowledge, identify and learn about its plethora of capabilities. The impacts of this technology may be incredibly far reaching, to which Raghavan adds “For a technology as profound and far reaching, the critical aspect is open participation. The key is to make sure that all of the elements, the creation, adoption, evaluation, benchmarking and safety of the model, is then done in an open and transparent fashion. That is the open ecosystem for AI that we’re talking about.”