Artificial intelligence is probably already being used somewhere in your organization. Whether it’s a chatbot or machine learning algorithm, AI programs need governance, discipline and ongoing care and maintenance to perform well, reduce risk and drive expected outcomes.
The three building blocks of AI? Data, algorithm and output.
Do you know:
It’s critical to consider how AI affects your business from a financial, operational and customer perspective and how it impacts your reporting needs.
Here's what our risk and assurance services include:
Using the defined PwC Risk Framework, we provide management with a readiness assessment report, sharing observations and recommendations to enhance your AI program and prepare for future attestation reporting. We also design and can implement specific systems controls for all aspects of AI (e.g., process oversight, data, models, output, machine learning), where permissible.
To help ensure performance meets expected goals we can perform AI testing. We review your models against documented policies and standards, and expected range of model outcomes. We also provide recommendations for tuning.
Our continuous monitoring services provide consistent oversight for risks in algorithms, minimizing disruptions and helping to mitigate risk, while also providing reporting for all stakeholders.
We can help build a robust AI governance and oversight program to give you a coordinated and disciplined approach to AI across your organization.
Through detailed testing of outputs against identified key risk domains, we can highlight potential bias in algorithm performance and provide recommendations for corrective actions.
Providing an independent, third-party perspective, PwC supports important new AI development, application and use, and can help to accelerate acceptance and adoption of critical new trusted uses of AI through third-party assurance services.
By incorporating AI processes into existing or new controls reports, management can obtain third-party assurance over design and operating effectiveness of controls that incorporate AI.
Managing Director, Assurance, PwC US