Should Generative AI be Open Source or Proprietary?
There are very strong opinions out there on how generative AI (genAI) should be developed, used and governed, and one of the key issues is about whether it should be open and free to be customized, or if its code should remain proprietary and available only with a fee.
The heated controversy over the topic has taken center stage in the media and among tech industry luminaries.
Listen the summay now: This Audio Was Generated with AI
One camp claims that users benefit from open source and developers who are able to provide continuous improvement. On the other side of the debate are those who believe that keeping AI proprietary can offer a safeguard to AI dangers and abuse by cybercriminals and bad actors. Both sides have a point.
This is a topic Wovenware cofounder, Carlos Melendez unpacked in an article for the Forbes Technology Council. To better understand the debate, Carlos says that it’s important to understand the difference between open source and proprietary AI. Open source is an AI program freely distributed for public consumption for users to build upon and share.
Closed, or proprietary, AI keeps its code under wraps, preventing its users from customizing it. As an example, Meta’s Llama 2 model is available as open source for the public to download and modify, whereas solutions from OpenAI and Anthropic both offer closed models.
The Benefits
Open source generative AI offers several compelling advantages over closed AI systems:
Collaboration and Innovation. Open source AI encourages collaboration and fosters innovation by sharing the collective expertise of researchers, data scientists, developers and engineers.
By sharing code, datasets and experiences, the performance and capabilities of models can be continuously improved and best practices shared to optimize the ways that they are used.
Customization. By keeping generative AI models open source, organizations, as well as developers, have the freedom to customize the model to meet the needs of specific applications, markets and users. They’re able to modify code, fine-tune parameters, and integrate additional features as needed.
Transparency. Open source AI models allow researchers, developers, businesses and end-users to have access to the code, algorithms, and data used to train the model. This transparency fosters trust and accountability in the AI system and enables any biases to be freely identified and addressed.
Despite its benefits, one of the biggest concerns of open source is the ubiquity of it. When it’s freely available there’s no way to control its use by “bad actors,” who can leverage the models to spread misinformation, produce deep fakes or other malicious activities.
Despite some caveats, one thing everyone can agree on is that open source genAI is revolutionary and quickly becoming pervasive across all industries and sectors of society. Yet, as with any such disruptive shift, it can take a village to ensure its most effective and safe use.
Related Content
Many Roads Can Lead to a Career in Data Science: A Conversation with Maxar ‘s Olivia Wendler
Many Roads Can Lead to a Career in Data Science: A Conversation with Maxar ‘s Olivia Wendler We recently sat…
Wovenware Moves Headquarters in Puerto Rico to Heart of San Juan’s Business District
Maxar Intelligence company, today announced that it has moved its local headquarters in Puerto Rico to a new location in…
The Wovenware Monthly AI Index September 2024
A Snapshot and Summary of What’s Happening to Enable the Safe and Effective Use of AI Welcome to the latest…