As artificial intelligence technology continues to advance rapidly, the ethical concerns surrounding its development and use are becoming increasingly apparent. In a recent interview, Igor Jablokov, the CEO and founder of AI company Pryon, discussed these pressing issues.
Jablokov highlighted numerous critical ethical challenges in AI, numbering around 20 or more. These include potential problems such as AI generating false information, compromising data privacy, leaking intellectual property from training on proprietary data, and being vulnerable to biases and harmful content in training data. He also expressed concerns about security vulnerabilities like embedded agents and prompt injection attacks, as well as the energy consumption and climate impact of large language models.
Pryon started from Jablokov leading an AI team at IBM. They worked on an early version of what became Watson. But IBM didn’t approve the project, so Jablokov left and founded Pryon. Later, Amazon bought Pryon, which led to creating Alexa.
RehumanizeNow, Pryon focuses on solving ethical issues with AI through responsible design. This is important for critical systems and high-stakes scenarios. The company lets enterprises control their data environments to protect sensitive data. Pryon also enables explaining AI outputs and tracing knowledge sources. This allows humans to verify the information AI provides.
Pryon uses supervisor approval before providing AI advice to workers. This ensures guidance is checked before reaching employees in energy, manufacturing, and healthcare fields. Oversight verifies results prior to technician delivery.
RehumanizeFor responsible AI growth, Jablokov promotes new regulations. Though he views the White House order positively, more action is needed against risks from generative AI technology.
While AI grows stronger, Jablokov highlighted the necessity of knowing the subject well and having emotional intelligence, especially in important decisions like healthcare where human judgment and understanding are vital.
RehumanizeAt the AI & Big Data Expo soon, Pryon will showcase fresh enterprise uses across industries like energy, semiconductors, pharmaceuticals, and government sectors. Jablokov suggested new ways to access the Pryon platform might emerge, potentially offering developers lower-level access paths.
Pryon’s approach to AI development emphasizes important aspects like governance, reliable information sources, human supervision, and collaboration with regulators. This model could set an example for responsible AI implementation across various industries as the technology continues to advance. By prioritizing these elements, Pryon aims to ensure AI systems are developed in a thoughtful and accountable manner.