By Jamie Bredehoft
With their artificial intelligence (AI) investments maturing, organizations in every industry are looking to maximize value from current data management practices. And, according to recent KPMG research, healthcare, financial services, and retail executives hope to pursue even more aggressive AI investments – but remain apprehensive due to emerging regulation, flaws in cloud environment security, and questionable integrity of data-driven insights. That’s where blockchain comes in.
“Blockchain’s core value proposition is to provide a chain of custody for data with trust as its key output,” says Tegan Keele, U.S. Blockchain Program Leader at KPMG. “Blockchain can also show where data came from, who accessed it and which models leveraged it, which ultimately helps mitigate data bias and validate conclusions and insights driven by AI.”
KPMG’s recent blockchain predictions for 2020 suggested that the convergence of these two advanced technologies would increasingly allow for blockchain to build trust in AI, and vice versa. And now the U.S. government has awarded KPMG a patent for a new method to make it happen.
“When blockchain and AI are implemented together as part of broader digital transformation efforts – versus adopted in siloes – organizations can be more transparent about how sensitive data is being controlled, and build new levels of trust among their key stakeholders,” Keele says.
On the heels of KPMG’s recent patent announcement, here are four use cases demonstrating how converged AI and blockchain adoption can drive value for organizations across industries. While AI may already be in use among the first two examples, leveraging blockchain’s additional trust layer to manage sensitive underlying data can maximize AI’s value:
Healthcare and Life Sciences Ecosystems. AI presents pharmaceutical manufacturers with an opportunity to identify predictive trends in supply and demand, and even adjust shipping routes or inventory levels in response. AI is also being explored as a way to track and analyze research from clinical trials to better predict patient drop-out or where participants may not adhere to clinical trial standards. Payers and providers may leverage AI to predict health outcomes in order to better calculate risk and adjust premiums accordingly. Pharmaceutical inventory and patient health data, while extremely sensitive, is extremely critical to enabling these types of AI-based predictions and responses. Blockchain can better secure and control the data, models and insights in order to prevent bad actors access, and validate any new data prior to using it to drive conclusions about patients or products.
Internet of Things (IoT) Device Security. Optimizing the value of IoT data with AI is nothing new. In manufacturing, AI can be used to better predict machine failures or other events impacting product quality. For IoT-enabled medical devices, AI may be able to better predict patient health events, enabling preventative intervention. Among autonomous vehicles and drones, AI is used to optimize everything from vehicle routes to settings based on weather conditions. However, IoT devices are inherently vulnerable from a security perspective. The addition of blockchain not only validates the actual IoT device but also enables data traceability and verifies AI model outputs, ensuring predictions, recommendations and adjustments are based on trusted data.
A second set of use cases emerges among maturing blockchain systems, where AI is valuable, but the need for data security and lineage is critical:
Telecommunication and Energy Settlements. We are increasingly seeing organizations derive value from implementing blockchain to settle financial transactions faster and with significantly reduced reconciliations. With 5G emerging, data exchanges will take place at a more rapid pace than ever before, with both the telecommunication and energy trading industries subjected to a significant shift in how they must operate. Reliable AI models that can predict volume or suggest optimal pricing adjustments will be essential to effectively managing this “new normal” and proactively combat rising costs to consumers. Organizations with existing blockchain networks that manage transaction settlements will be in a much better position to realize these benefits from AI, and can even share models or insights with select partners.
Retail Loyalty Programs. Blockchain-based tokenization offers a new form of value exchange within loyalty networks, allowing consumers to use points for purchases with different merchants. With blockchain in place, loyalty program providers will soon be looking to AI to personalize the program while optimizing the cost to operate. AI can be introduced to predict consumer behavior, incluing what they'll buy, when and where, as well as offer dynamic reward suggestions. The "price" of a reward can also be adjusted using AI based on conclusions about a consumer’s spending habits, or from supply and demand changes resulting from broader market events.
While insights based on AI can be, and have historically been, achieved without blockchain, trusting the model and the ability to explain why a certain insight was generated can be a major challenge. Turning to blockchain can help organizations operationalize a trust paradigm within their AI lifecycle and optimize AI’s value by building and retaining the confidence and trust of internal and external stakeholders.
To learn more or to arrange an interview, please contact Jamie Bredehoft..