The 5 Questions to Ask Before Purchasing AI

Posted by Kevin Thomas on December 21 , 2018

The 5 Questions to ask before purchasing AI

1. Can you identify the risks that the technology addresses?

2. Can you ensure accurate data inputs?

3. How will the technology affect stakeholder relations and security practices?

4. Can it be tested to ensure transparency and accountability?

5. Will the technology be used in a manner that respects the autonomy of the people it will affect?

Keep reading to see how Vintra helps you answer these vital questions.

In a recent edition of the Technology Quarterly by The Economist, the authors investigate what the advancements in technology, and specifically Artificial Intelligence (AI), mean for the citizen, law enforcement, and businesses. In the report, they quote author Andrew G. Ferguson who posed five key questions that we have modified and believe are worth companies and agencies answering before they purchase the latest-and-greatest in AI. 

1. Can you identify the risks that the technology addresses?

From the minuscule to the grand, Vintra's video analytics addresses the risks and concerns related to safety that every agency and modern enterprise must account for.

  • In post-event situations, manually reviewing video is costly and inefficient. In a world where 98% of surveillance video is never viewed, the need for a solution to accurately review and monitor the video, and make actionable, is extremely urgent. When security officials are able to sit down and review video, it is done, unfortunately, incredibly ineffectively. That's because after only 22 minutes of sustained review or monitoring humans experience something akin to video blindness, and miss 95% of activity on the screen. As the quantity of video continues to explode in amount and variability of sources, without the proper tools to handle the volume and detail of review required, operators and investigators are left fighting an uphill battle, at best.
     
  • For real-time situations, the risks of of maintaining status-quo in regards to safety and security monitoring
    increase in severity every year. Violent crimes and active shooter incidents in the workplace and/or public gathering places are on the rise. Relying on security guards or central monitoring stations staffed with well-meaning, but unequipped humans, to react quickly and proactively stop these horrific events before they start is like plugging holes in a raft with your fingers. On the business justification front, gains in productivity, decreases in worker's comp losses, a reduction in compliance standard violations, and vastly more secure entries and exits that contribute to a reduction of data breaches are all left on the table without a comprehensive video analytics solution. The ideal solution provides actionable intelligence that can be implemented in the enterprise and security sides of any business or law enforcement agency. 

2. Can you ensure accurate and secure data inputs, and the security of the entire workflow?

  • Yes. Your video or live video streams are never altered or manipulated in any way by FulcrumAI. We are a powerful tool for better search, collaboration, and intelligent-monitoring of any type of video. In real-time settings, the solution connects directly to the cameras or VMS and is not routed off-site in any way. The same security and access standards that apply to your VMS apply to the integrated use of FulcrumAI. For the post-event cloud solution your data is encrypted on the way up and down, user logs can be tracked and accessed by administrators, and the original video is never touched by Vintra. 

Read the latest Vintra White Paper, "The AI (R)Evolution of Enterprise Security"

Get the White Paper

 

3. How will the technology affect stakeholder relations and security practices?

  • It is our hope that the those who implement our technology would use it to foster a greater sense of trust within the various communities they serve, be they public or private. In the current social and political climate there is much focus on the ways in which enterprises and law enforcement agencies engage with the public. There are three ways in which our technology can help promote trust among stakeholders, as well as improve security practices:
    1. Video Allies: a common frustration is that incident resolution is slow, if at all. In many cases, this is due to a dearth of video evidence, as well as a lack of ability to collect critical video from 3rd parties that may be out there, and prove helpful to both public and private agencies. Through our Video Allies feature, it is now possible for the community and 3rd party entities to upload their video directly into the solution so that it can be processed and searched immediately. Imagine how much video an incident like the Boston Marathon bombing or a shooting at large entertainment complex would elicit from the community, and how much quicker evidence could be acted upon. Not only does the feature empower enterprises and public agencies to collect and search video far more efficiently, but to give the community a sense of involvement in the resolution of said incident.

    2. If the community was made aware of an agency's use of FulcrumAI, then they would also be more confident in the results of video-driven investigations and evidence since the bias and physical limitations of humans when it comes to video review is, as much as possible, eliminated from the equation by way our solution. As an employee of Company X, you could also be confident that the entrances, exits, and critical areas of the office are more secure — leading to more personal security — if Company X employs facial recognition to augment their badge access measures. 

    3. The company or agency can also use their partnership with Vintra, and our commitment to the fair and unbiased training of our facial recognition algorithms, to lend credence to their desire to only use face rec when and as necessary. And not, as is widely feared, to maintain a constant and indiscriminate knowledge of each person and their actions within reach of the agency or company's jurisdiction. 

Watch this 90-second video to see what's possible with FuclrumAI

See What's Possible

 

4. Can it be tested to ensure transparency and accountability?

  • At the agency's request, Vintra can provide a report of use that can verified by a neutral third party who has been appointed by the legal stakeholders of the company or public safety agency to confidentially review the accuracy of results in cases which FulcrumAI has been used. Due to ongoing cases and privacy concerns, a subsequent report can then be produced and shared with the appropriate legal stakeholders for further review.

5. Will end users deploy the technology in a manner that respects the autonomy of the people it will affect?

  • The technology will primarily be used in post-event situations in which any person, vehicle, or identifying feature of someone present in video is searched for only after an incident has been reported, and to assist in resolving the incident quicker or more accurately. 

  • In real-time monitoring instances, our technology is used primarily by private entities to better secure their physical spaces and optimize their business operations. In such cases, the employee's privacy is ensured by the company's employee privacy policy. In no case should the company use our technology to perform any kind of analysis or data collection on their employees with their consent. 

Subscribe to the blog by filling out the form below.

 

Topics: Artificial Intelligence, AI, Physical Security, Security, Public Safety, facial recognition, Deep learning, Enterprise Security, White Paper, AI Conference, O'Reilly AI Conference

Comments are closed.