Enterprise Use of ChatGPT and Alternatives

We are all talking about?ChatGPT, the chatbot launched last year by?OpenAI, in personal as well as profession conversations. At least once a day, I see it come up in a context, or from a person, that I didn't expect.

My colleague Dilip Bhatia wrote an engaging post about it recently where he explored the topic of Customer Experience. I raised a point about my concerns with Enterprise use of ChatGPT and was asked to share my thoughts. So here goes.

My view, as stated in the comments to Dilip's post, is that the challenge for us in Enterprise (as opposed to personal) space is working out the implications (legal, augmentation vs replacement, speedy but general responses vs nuanced, customized ones - to specific customer situations, and so on). When asked to share my thoughts, which I am doing here, I decided to follow Dilip's lead and ask ChatGPT.

Question: How can enterprises use ChatGPT in a legal and transparent way?

ChatGPT Answer: Enterprises can use OpenAI's GPT-3 (which includes ChatGPT) in a legal and transparent way by following these steps:

1. Review OpenAI's terms of service: OpenAI has specific terms of service that enterprises must follow, including guidelines for ethical use and compliance with all relevant laws and regulations. It's important to thoroughly review and understand these terms before using the technology.
2. Obtain necessary permissions: If the enterprise intends to use GPT-3 for commercial purposes, it must obtain the appropriate license from OpenAI.
3. Be transparent with end-users: If the enterprise intends to use GPT-3 for customer-facing applications, it should be transparent about the technology's use. This includes disclosing the use of AI, and obtaining the appropriate consents from customers if required by law.
4. Ensure compliance with data protection laws: If the enterprise intends to collect, store or process personal data using GPT-3, it must ensure compliance with all relevant data protection laws, such as the General Data Protection Regulation (GDPR) in the European Union.
5. Conduct regular audits and evaluations: The enterprise should regularly evaluate its use of GPT-3 to ensure it remains compliant with OpenAI's terms of service and all relevant laws and regulations.

By following these steps, enterprises can use GPT-3 in a legal and transparent way while leveraging its capabilities to improve their operations.

My thoughts: This is a detailed, albeit a bit verbose, response that makes a lot of sense. One of my research interests is Trustworthy and Responsible AI and transparent use of AI is one responsible way to earn a user's trust. Spot on, ChatGPT!

A follow-on question popped up in my mind as I continued the dialog with ChatGPT. The thought was along the lines of enterprises avoiding the use of ChatGPT by setting up and hosting their own alternative technology.

Question: Should Enterprises train and host their own large language models like GPT, for commercial use, instead of using ChatGPT?

ChatGPT Answer: No, it's generally not recommended for enterprises to train and host their own large language models like GPT for commercial use. Here's why:

1. Cost: Developing and training a large language model like GPT is a complex and resource-intensive process that requires a lot of computational power and data. This can be expensive for companies, especially small to medium-sized enterprises.
2. Maintenance: Once the model is trained, it requires ongoing maintenance and updating, which can be time-consuming and requires specialized skills and resources.
3. Expertise: Building and training a large language model requires a significant amount of expertise in deep learning, natural language processing, and large-scale data processing, which most companies do not possess in-house.

For these reasons, it's more cost-effective and practical for most enterprises to se pre-trained models like OpenAI's ChatGPT, rather than trying to build and maintain their own models. This also ensures that the model is always up-to-date and performing optimally.

My thoughts: Again, a well-crafted response that lays out the key drivers in favor of using ChatGPT. I do still continue to deliberate on the pros and cons.

Your thoughts: What do you think? Do share your thoughts as this is going to be a key decision every enterprise will need to make soon, if they haven't already done so.

P.S.: There is a lot of excitement around Large Language Models like ChatGPT. Check this alternative, called Bard, announced by Google today.