Embracing Generative AI From a Product Development Perspective
As the world of generative AI continues to evolve at an astonishing pace, businesses must stay informed and test new solutions, while also being cautious of how it's being used within the organization. With the rapid expansion of AI tools, it's essential to ensure that your team is using these resources responsibly by protecting trade secrets, consumer privacy, and maintaining accuracy in content generation. Here at bluesalve, we aim to guide you through the complex landscape of AI while keeping your organization's best interests in mind.
The Rise of Generative AI
Over the last six months, generative AI has become a vital tool for many in automating daily tasks. However, AI tools like ChatGPT can often provide incorrect answers and lack important context on niche topics. How can you be certain that the AI services or apps that your team uses are trustworthy and won't compromise your company's information or trade secrets?
With new advancements announced daily, it can be challenging to stay informed. Last month alone saw over 30 significant announcements and product launches. Each one could make up its own lengthy blog post. Amazingly, there are over 100 Large Language Models (LLMs) in use today, all competing to be the best. We have been tracking all the developments and the almost-daily research papers that show new ways to build upon these LLMs, or chain them together.
Some key developments that we are actively testing or reviewing APIs include ChatGPT Plugins, Whisper API, and ChatGPT API, as well as new enhancements built upon the LLMs, such as HuggingGPT, Microsoft Jarvis, and AutoGPT. We see great potential in how some of these developments can expand our client’s future product roadmaps.
Generative AI's Role in Product Development
If you look across your teams, you will see that AI has the potential to permeate every aspect of product development, if it hasn’t already done so. Every role, from marketing, product managers, engineers, and UX/UI, to finance and legal, all can use AI.
Whether established products like Notion AI, Figma AI, Adobe FireFly, Canva AI, Hubspot ChatSpot, Github copilot, or the vast Chrome extensions, there have been over 1,300 AI apps introduced recently. So it becomes crucial to scrutinize their use to ensure the protection of proprietary or non-public data. Here are some things to think about as you or your team test these solutions:
Safeguarding Confidentiality and Ensuring Accuracy
Anytime you enter information into the chat window or use an online Chrome extension to upload your PDF, or even use a prompt template, there is an opportunity for these tools to capture your private information. Some T&Cs are scary to read through, and sometimes they’ve even used ChatGPT themselves to create boilerplate terms.
When it comes to content creation for copy, the potential risks associated with AI-generated content cannot be overlooked. For example, creating content for a health wearable product requires accurate information and legal review, as reputational stakes are high if inaccuracies emerge. The key is approaching AI as you would in a large enterprise, implementing proper structures for success and growth. Even if you are a startup with just a few employees, taking the time to spell out clear policies can save you many potential headaches in the future.
Navigating Your AI Digital Transformation Responsibly
To harness the power of AI while maintaining data protection and accuracy, consider these four elements as a framework for responsible AI with your business:
- Principles: Work with team leaders to establish core values such as transparency, data governance, privacy and security, and inclusiveness.
- Practices: Develop guidelines for everyone that clearly spell out what content can or can’t be used with these tools. (Good practice: don’t enter proprietary or non-public data). Whether the AI interface is built into an established app, Chrome extension, or any other method out there. It’s important to work with your teams to use tools responsibly and create checklists to put these principles into action.
- Tools: Implement best practices across the entire organization. Establish an internal review process for the AI tools that are used, and to review the Terms & Conditions, and Privacy statements.
- Governance: Ensure C-suite / Owners of the business oversight; promote responsible AI leadership and the establishment of Responsible AI Policies.
As your trusted guide in the world of Artificial Intelligence, bluesalve can help you navigate this ever-changing landscape. Stay tuned for our next blog post as we discuss how your business should prepare your data to work with LLMs or build upon them with your own embeddings that are fine-tuned to your niche. By acting now and implementing key foundational steps, you can protect your company's reputation and trade secrets while unlocking the tremendous potential.
# # #
Rich Bira, SVP of Product Innovation, is actively engaged in the AI space and a committee member of Consumer Technology Association’s AI In Health Care and General Principles of AI/ML Standards Committees.