The White House has released 10 principles for government agencies to adhere to when proposing new AI regulations for the private sector. Karen Hao and MIT Technology Review does a good job of summarizing these principles here. The following is that top 10 as well as Lucd’s view on how enterprises should think about that principle.
Public trust in AI. The government must promote reliable, robust, and trustworthy AI applications.
Lucd view: Completely agree. It is easy to replace “government” with “enterprises” and restate the sentence. Enterprises must promote reliable, robust, and trustworthy AI applications. AI could not and should not be a skunk works initiative or a black box. AI innovation and development needs repeatable, reproducible, continuous innovation. That is why Lucd is investing in a transparent and governed AI innovation platform for businesses.
Public participation. The public should have a chance to provide feedback in all stages of the rule-making process.
Lucd view: This is critical for businesses as well. Enterprise AI is not a data science project. Multiple people and organizations in a business need to participate and have a chance to provide feedback in all stages of developing the AI powered solution. That is why Lucd is investing in a modern 3D multi-party UI for the AI development process so that many disciplines in the organization can participate.
Scientific integrity and information quality. Policy decisions should be based on science.
Lucd view: This mindset is required for businesses to compete in the age of AI. Basically this is what is means to be a data driven organization. Just like in science, the innovation should rely on data. But, also just like in science, the data needs to be able to be constantly questioned, evaluated, added to, and reassessed. That is why Lucd is investing in its secure, compliant Unified Data Space.
Risk assessment and management. Agencies should decide which risks are and aren’t acceptable.
Lucd view: This goes back to AI participation throughout an organization. AI cannot be a black box. People throughout an Enterprise need to understand how AI works and be able to question so that they can also understand the risks and rewards. Collaboration is critical to AI success.
Benefits and costs. Agencies should weigh the societal impacts of all proposed regulations.
Lucd view: This is an important point. Many people discuss whether it is healthy or not for AI innovation to be dominated by a few large tech companies. That debate will continue. However, Lucd believes that every business can and should be an AI business and be able to provide better products and services to its customers by leveraging AI responsibly to learn from its unique data. That is why Lucd works to make the economics on Enterprise AI work for a business with an efficient cost effective capability.
Flexibility. Any approach should be able to adapt to rapid changes and updates to AI applications.
Lucd view: Any business that invests in a proprietary approach to AI, even if it is the best approach today, needs to understand that there is likely somebody somewhere building a better approach right now. That is why Lucd is committed to open standards and the ability to support any open source or third party AI model to be trained on a businesses data. This is the definition of AI Democratization…as it enables organizational control and independence.
Fairness and Nondiscrimination. Agencies should make sure AI systems don’t discriminate illegally.
Lucd view: If AI systems are not developed with explainability, lineage, and governance, then assurance that there is illegal discrimination becomes very difficult to prove. Lucd has built an AI platform with explainability, lineage and governance. As a result, it becomes easier for government agencies, departments or businesses to ensure the data provenance and integrity.
Disclosure and Transparency. The public will trust AI only if it knows when and how it is being used.
Lucd view: As stated above, transparency is key. No Enterprise AI application should be developed in a “black box” approach.
Safety and Security. Agencies should keep all data used by AI systems safe and secure.
Lucd view: This is the cornerstone to trustworthy and successful AI. No Enterprise AI platform invests as has the security and compliance capability that the Lucd Unified Data Space provides.
Interagency Coordination. Agencies should talk to one another to be consistent and predictable in AI-related policies.
Lucd view: AI collaboration across agencies or business units is critical. A Robust Enterprise AI solution needs to incorporate the modern collaboration capabilities. These dynamic capabilities are available today. By including multi-party gaming, version control systems, and the ability for non-data scientists to participate, true AI collaboration can be achieved.