Much is said about algorithms in artificial intelligence discriminating against different groups of people. Whether this is intentional or unintentional it poses serious problems that can ruin a company’s reputation permanently. This can also overlook many potential options unconsciously denying a business of a preferable choice.
Preventing Discrimination and Encouraging Fairness
Amazon was using an AI tool that was later found to discriminate against women and therefore had to do away with it. Apple’s new credit card is being accused of discriminating against sexist lending models. Transparency among AI models is necessary and will help to solve some of these issues and gain the trust of the public.
Security Risk with AI Transparency
With AI comes convenience and at the same time, with AI transparency, comes an open door to hackers. AI transparency means opening up and sharing the methods by which they create their algorithms. Disclosure can not only open explanations to hacking but it can also open employers up to even more regulation.
Companies also become more susceptible to lawsuits which are not the goal of transparency. Transparency is for the protection of all and not to create more government regulation but to create accountability and fairness. Here lies the AI transparency paradox, creating an open and honest forum is good and necessary but this open forum creates complex problems for companies.
Releasing the fine details about an AI’s algorithms only increases liability and risk, but then again, all data carries its risk. As AI is accepted more and more and becomes mainstream the more downfalls, bugs, viruses, etc. will be discovered.
AI Has a Long Way to go in Development
There is still a lot to learn concerning artificial intelligence and how we can simultaneously be fair and protect our programming, our rights, and our safety. The more information is released about an algorithm the more harm that a malicious hacker can cause. This is not to say that we should throw the baby out with the bathwater.
It is necessary for people to understand that with transparency, comes risk. As AI is progressing so is AI security, an entire network dedicated to catching and exposing risk before they cause harm.
Lastly, another form of protection is to employ lawyers as often as possible when dealing with AI. Lawyers create a legally privileged space, allowing companies to thoroughly decipher and look for vulnerabilities in their AI models without the worry of more liabilities and unwarranted trouble.
~ Written for us by our associate Gary Sorrell, Sorrell Associates, LLC. Copyright protected. All rights reserved.
No comments:
Post a Comment