As new technologies incorporate artificial intelligence (AI) and overall uses of AI continue to expand, what protections are in place to prevent its misuse? Artificial intelligence, like any technology or security measure, is not inherently bad. It can, however, be used for nefarious purposes and in ways that were never intended or thought of.
‘Responsible AI Policy’ to protect users
Prosegur has created a ‘Responsible AI Policy’ to protect users from potential AI threats
As an AI company, Prosegur has created a ‘Responsible AI Policy’ to protect users from potential AI threats, and to prevent employee and partner misuse of AI-enabled technology, with guard rails in place to hold them accountable, if artificial intelligence is being used nefariously.
“Any security or similar company that utilises AI with its products should monitor for potential misuse and implement a similar policy for protection,” contends Mike Dunn, the Chief Technology Officer (CTO) at Prosegur USA.
Requirements of Prosegur’s ‘Responsible AI Policy’ include:
- Human action and oversight - Prosegur heavily monitors their agents using an integrated international security operations centre (iSOC). Floor and shift supervisors ensure that everything is working properly and that incidents are handled. They also ensure employees are logging into AI-enabled cameras and other technologies for the right reason.
- Transparency - As with any new company-wide policy, Prosegur wants to ensure that every employee learns about the policy and their duties to uphold it. “In doing this, we are making sure that all processes involving AI development are transparent, so any employee can understand and implement the policy into these processes,” says Mike Dunn.
- Non-discrimination and equity - Prosegur seeks to make sure employees are using AI-enabled technology to look for known criminals, for example, and not profiling or singling people out. Having a policy in place ensures agents are aware.
- Accountability - The most important part of the policy is to hold people accountable, if AI-enabled technologies are misused. Violation of the policy will result in disciplinary actions, which can range from HR (Human Resources) intervention to possible suspension. For partners, Prosegur would cease and desist using their products immediately upon finding out about any misuse.
Leading the charge for ethical AI practices
Prosegur is proud to lead the charge for ethical AI practices, by creating this first policy of its kind"
“Prosegur is proud to lead the charge for ethical AI practices, by creating this first policy of its kind, and we hope that other security organisations follow suit,” says Mike Dunn.
Technology can do wonderful things. In security, technology can help to stop crimes and even to prevent them from happening with deterrents and alarms.
Important to take action to stop ‘bad actors’
However, as with any industry or group of people, there will be ‘bad actors.’ Mike Dunn said “It is important to take actions to stop those people, but it is also important not to condemn all technology or practices when this happens.”
Even proven, effective technologies can be used for bad (as well as good). An example is TSA Scanners at the airports. They were designed to provide safer air travel, picking up things that normal metal detectors could miss.
Misuse of AI technology can cause lot of harm
However, it came to light that a few ‘bad actors’ were using this technology and sharing silhouetted nude pictures of passengers. This has since been patched and fixed, but nonetheless, it’s a great example of how misuse can break people’s trust.
As a whole, technology can be dangerous. If not used as intended, it can be used to hurt people. The U.S. government has found that China was using facial recognition and demographic analytics to classify and identify people, based on their ethnicity and that technology singled them out for tracking.
Important to ensure analytics and AI used as intended
As a society, I think it’s important for us to take that first step and self-police any use of these analytics"
“As a society, I think it’s important for us to take that first step and self-police any use of these analytics,” said Mike Dunn, adding “At a minimum, (we should) set policies and procedures to make sure these analytics and AI are being used as intended.”
As use of artificial intelligence (AI) continues to increase, an industry standard would ensure companies have protections in place for their employees, clients, and partners. Having a policy is the first proactive step in ensuring AI protections. However, companies need to enforce accountability too.
Effectively handle security threats and vulnerabilities
Mike Dunn continues, “By heavily monitoring for unethical AI practices and enforcing disciplinary actions when needed, companies will be better prepared to handle security threats and vulnerabilities.”
At Prosegur, monitoring agents are heavily trained, taught procedures and then closely supervised by several methods. “This is something that should be common practice in the security industry,” says Mike Dunn, adding “We have employed a compliance officer to help us stay ahead of the curve. Fortunately, we have seen many manufacturers and end users adopt this practice as well.”
Compliance officers can act faster than new laws enacted
Compliance officers can act faster than new laws being enacted, and often will hold their company to higher standards than a minimum threshold.
Mike Dunn concludes, “Putting a responsible AI procedure in place is Prosegur’s way of taking on this problem we see in our everyday work.”