Responsible Artificial Intelligence
Verizon has developed principles to guide its AI development, acquisition, deployment, and use.
Verizon, like many companies, uses AI to improve our products, services and business operations. We understand that poorly governed AI applications can result in unintended consequences, including potential bias or discrimination, whether in design, implementation or the data sets used to train AI models. Verizon’s Artificial Intelligence and Data organization, together with our Data and Analytics Office, manages and implements Verizon’s centralized, enterprise-wide Responsible AI Program.
Our Responsible AI Program is overseen by a dedicated team with support from other internal subject matter experts, including our Privacy team and Business & Human Rights Program. This program is governed by a risk-based approach and informed by emerging global definitions, concerns, frameworks, regulations and legislation in AI and related fields.
We are guided by a set of foundational principles that provide direction to the many V Teamers using AI and align Verizon’s use of AI with the values of our Credo, as summarized in the table to the right. While these principles do not represent the full breadth of AI issues, they are intended to inform common terminology for AI-related efforts, guide our efforts to leverage new AI technologies in ways that positively impact our stakeholders and establish Verizon as a trusted brand and partner with respect to AI.
Verizon’s Responsible AI Principles
Governance | Verizon employs a risk-based framework for developing, using and overseeing AI. We have processes to oversee Verizon’s development and use of AI, and our highly skilled employees have an important role to play in considering the risks, impacts and benefits of using AI. |
|---|---|
Respect for privacy | Verizon recognizes that AI systems must incorporate a privacy-by-design approach, including consideration of the data used to train models and the data used to make business decisions. Verizon’s use of AI must comply with our privacy policies and applicable laws. |
Respect for human rights & responsibility to society | We design and train Verizon systems using strategies to identify and reduce potential bias or other harms. We address the risk of unlawful or unintended bias in AI systems with appropriate governance and mitigation measures. |
Technical robustness | Verizon uses industry-accepted metrics to evaluate AI solutions’ ability to generate accurate, reliable and reasonably consistent outputs. We employ a security-by- design approach, which proactively and holistically monitors for and mitigates against security threats. |
Transparency | Verizon understands that in order to provide reliable and positive user experiences, our use of AI systems requires transparency. Verizon is committed to clearly explaining its governance, use and monitoring of AI systems. |