- In the seat this week: Ryan Steelberg, president and CEO, Veritone.
- About Veritone: Veritone designs human-centered AI solutions. For law enforcement, its aiWARE platform and iDEMS public safety suite can reduce administrative workload, especially around video/audio evidence, public records requests and investigations.
- About Steelberg: With over 25 years of executive management experience in technology, marketing, business development and sales, Steelberg has formed numerous companies in the tech space, including cofounding Veritone in 2014. Since 1996, he has raised several hundred million dollars in equity financings, managed more than 1,000 employees and created several billion dollars in shareholder value through the successful founding, scaling and liquidation of multiple industry-defining companies.
- Learn more: Veritone launches AI-powered iDEMS for public safety and judicial agencies.
One of Veritone’s guiding ideas entering 2026 is the idea of “data as currency.” What does that mean, and how does it inform what you do?
“Data as currency” highlights the value of unstructured data – including video, audio and digital files – and its potential to drive economic and practical returns when refined and utilized effectively. At Veritone, this principle drives strategies like data tokenization and monetization through our AI platform, aiWARE, and Veritone Data Refinery (VDR), turning raw data into actionable tokens embedded in workflows. This process positions us as leaders in the data economy, preparing for a future of rapid AI-driven expansion, including the projected $13.55 trillion tokenized asset market by 2030.
Where do you see the most immediate value opportunity for AI in law enforcement today and why?
The most immediate value for AI in law enforcement lies in revolutionizing evidence management and solving crimes faster. AI can process vast amounts of unstructured data far faster and more accurately than humans, enabling efficient review of audio and video footage (such as body-worn cameras, car-dash cameras, surveillance cameras, mobile phone and social media video clips, just to name a few), rapid identification of key events and swift transcription of critical audio files. By alleviating resource strain and streamlining workflows, AI allows agencies to accelerate investigations faster, increase case clearance rates, allocate resources more effectively and ultimately improve public safety while providing greater transparency.
What does “human-in-the-loop AI” mean in practical terms for policing? Where do you believe human review is essential versus where full automation could be appropriate?
“Human-in-the-loop AI” refers to a collaborative model where humans guide AI processes for accuracy, ethics and context-specific decisions. In policing, human review is essential for interpreting complex legal or ethical contexts, detecting potential biases and making high-stakes field decisions. Full automation, however, is effective for tasks like categorizing data, identifying faces or objects in surveillance videos and flagging suspicious activity for human review. This balance ensures AI enhances efficiency while maintaining ethical oversight and accountability. Part of Veritone’s Responsible AI policy is incorporating “human-in-the-loop” processes into our AI-powered software applications.
If you could influence national or state-level policy on AI in public safety, what standards, enhancements or guardrails do you think would most benefit agencies and the public?
National and state policies should prioritize transparency, requiring AI tools to clearly explain their decision-making processes and undergo regular audits to ensure fairness and eliminate biases. Privacy and data security standards must safeguard sensitive information, and interoperability between AI platforms should enhance seamless integration across agencies. Additionally, robust training programs are crucial to educate officers on AI tools, limitations and ethical implications, enabling responsible adoption that benefits law enforcement and the public.
With video, audio and digital files increasing exponentially, where do you see the biggest bottlenecks emerging? How is Veritone preparing for evidence environments five years from now?
The biggest bottlenecks are rooted in fragmented systems, data silos and the lack of interoperability across tools. This leads to issues with efficient data organization, accessibility and analysis speed. Traditional infrastructure struggles to keep up with the exponential growth of high-resolution evidence from diverse inputs like bodycams, drones and surveillance networks. Veritone is addressing these challenges by prioritizing an open architecture approach with its aiWARE platform and VDR. This focus ensures integration with existing systems, avoiding vendor lock-in and enabling agencies to unify their evidence workflows. By streamlining data tokenization, indexing and retrieval at scale, Veritone can handle the growing volume of digital evidence while future-proofing agencies against the risks of outdated, siloed technologies. This open and scalable design empowers agencies to optimize workflows, maintain flexibility and stay ahead in increasingly demanding evidence environments.
For executives reading this, what are the top misconceptions agencies have when evaluating AI platforms, and how can vendors better support chiefs through procurement and rollout?
A key misconception is expecting instant results without preparation, overlooking the importance of training datasets and workflow integration. Agencies may also fear AI will replace human oversight, causing resistance among staff. Vendors can address these concerns by providing clear, realistic expectations of AI capabilities, demonstrating its role in augmenting decision-making. By offering tailored demonstrations, ongoing training and hands-on support during rollout, vendors can build trust, ease transitions and ensure AI solutions are fully optimized for agency use.
Over the last couple of years, we’ve heard a lot about novelty and proof of concepts with AI. The organizations that will succeed in seeing value from their investments in AI are those that understand their data sets and then begin implementing AI to solve simple, manual workflows before scaling to meet their needs.
