The Five Most Key Takeaways from This Blog
- Open-source A.I. development has been a hot-button topic in the A.I. world for some time now. It refers to making the “source code” for certain A.I. projects openly available. That open availability translates to allowing people to freely modify, use, and even share their modified versions of the A.I.
- The United States government has agreed to let open-source A.I. to continue without restriction. However, this may change down the line.
- Calls for restriction come from many sources. Chiefly, there are people who believe that protecting A.I. secrets can keep bad actors from abusing and misusing these systems for nefarious ends. There are also corporations that appreciate keeping A.I. software a secret, much like how soda companies seal up their soda recipes as trade secrets.
- For businesses with technical know-how, this could allow for the creation of A.I. tools using existing open-source projects, some of which may be the projects of tech companies with a ton of resources for building sophisticated A.I. systems.
- IBM lists a number (five, the number is five) of open-source tools that business owners could potentially benefit from.
The Not-So-Secret Formula
If you are familiar enough with the culturally (American, but in many spots of the globe as well) iconic show Spongebob Squarepants, then you are likely aware of the coveted Secret Formula. Or, as Mr. Krabs would call it, “The Secret Formuler”.
That closely guarded recipe for the famed Krabby Patty has many parallels to the source code of many high-value projects in the world of technology.
Those projects can include A.I. projects.
Open-source A.I. projects do away with the Secret Formuler approach in favor of allowing the public to access the source code.
However, certain people recommend restricting open-source projects because of the potential for misusing A.I.
Below, we consider the pros and cons of open-source, along with the potential impact on business owners or restricting it.
The Impact of Open-Source Restriction on Business Owners
If you restrict open-source development, businesses that have the resources to modify existing open-source A.I. would have much less power in creating custom-fit A.I. that internal developers could create using pre-existing frameworks.
That is not to suggest that custom-fit A.I. would go away. No, not at all. Rather, the large tech companies that create A.I. would allow businesses to custom-tailor A.I. solutions after purchasing a subscription plan, or making a one-time payment for a custom project, or the like.
Basically, businesses would have less autonomy in participating in A.I. development, and instead need to rely on large tech companies that have the Secret Formuler to get the A.I. solution that they want to implement.
The other side of this, however, is that it may make it more difficult for certain businesses to make some off-market A.I. for its own ends, without the usual oversight that a larger tech company would be beholden to observing during development.
Why Do Certain Tech Companies Support Open-Source A.I. Development?
However, something that is pretty important to know about the tech world, specifically when it comes to the history of the Internet, is that there is an idealistic bent to a lot of tech workers’ worldviews.
For many such workers, that involves things like the “democratization of information”, wherein the common people (in contrast to, say, wealthy and powerful elites) have free access to information.
Here is an example: in the pre-Internet days, if you wanted to access specialized chemistry knowledge, you would likely have to go to a university library. And, likely, have to be enrolled in that university to access the books or papers. Post-internet, there are plenty of highly specialized scientific articles available to anyone with an Internet connection and Internet-connected device.
Something like this informs the approach of some tech workers–and, importantly, tech companies’ leaders–to A.I. development.
But it is not just idealistic principles that appeal to tech workers. For some, the main draw is that you can spur greater information when you allow members of the public to take a looksee at the source code and offer ways to help.
What’s more, you get free labor out of people who just want to contribute to the open-source project for the fun of it, or the prestige or recognition that doing so could earn them.
Why Do Certain Tech Companies Not Support Open-Source A.I. Development?
Well, to protect their Secret Formulers above all. If everyone was able to see how ChatGPT works (OpenAI may be the most notable example of the anti-open-source camp), then it would be much easier to compete with OpenAI then, wouldn’t it?
Other reasons are also cited by companies, such as observing responsibility. There is currency to the argument that these A.I. projects are markedly different from, say, an open-source photo-editing software.
These are powerful programs that, if misused, can cause plenty of harm, social or financial or otherwise.
