Human Involvement Can Help Prevent AI Mistakes

That’s what security experts say artificial intelligence (AI) systems used by companies can make serious and costly mistakes. But one way to avoid such mistakes is for companies to employ people to keep a close eye on the AI.

An example of AI issues that can impact businesses emerged at the start of the COVID-19 pandemic. This involved the credit scoring company Fair Isaac Corporation, known as FICO.

FICO is used as a guide by about two-thirds of the world’s largest banks loan Decisions. The Company’s systems are also used to identify potential credit cases fraud.

FICO officials recently told Reuters that one of the company’s AI systems misidentified a large number of credit card fraud cases. At that time, the pandemic had caused a sharp increase in online shopping. The AI ​​tool viewed the rise in online shopping as a result of fraudulent activity.

As a result, the AI ​​system instructed the banks to reject millions of purchase attempts from online shoppers. The incident happened just as people were rushing to buy products that were in short supply in stores.

But FICO told Reuters that very few buyers ended up being denied their purchase requests. This is because a group of experts that the company uses to observe or monitor its AI systems caught the false fraud detections. The workers made temporary adjustments to avoid a spending freeze ordered by the AI.

Also Read :  How executives can prioritize ethical innovation and data dignity in A.I.

According to FICO, the expert team is quickly notified of any unusual buying activity that the AI ​​systems may misidentify.

But these types of corporate teams are not that common, Reuters reports. Over the past year, FICO and management consulting firm McKinsey & Company have conducted separate studies on the subject. They found that most of the organizations involved in the study were not closely monitoring their AI-based programs.

Experts say that AI systems make mistakes primarily when real-world situations differ from the situations used to create the intelligence. In the case of FICO, it was said that its software expects more in-person shopping than online shopping. This resulted in the system identifying a larger proportion of financial activity as problematic.

Seasonal differences, changes in data quality, or extremely unusual events – like the pandemic – can lead to a range of poor AI predictions.

Aleksander Madry is director of the Center for Deployable Machine Learning at the Massachusetts Institute of Technology. He told Reuters the pandemic must have been a “wake-up call” for companies that aren’t closely monitoring their AI systems. This is because AI mistakes can cause big problems for companies that don’t do it effectively administer the systems.

Also Read :  Updates to Apple’s zero-day update story – iPhone and iPad users read this! – Naked Security

“That’s what’s really holding us back from this dream right now of AI revolutionizing everything,” Madry said.

Urgency has been added to the topic as the European Union plans to pass a new AI law as early as next year. The law requires companies to monitor their AI systems. Earlier this month, the US government also proposed new policies aimed at protecting citizens from the harmful effects of AI. In the guidelines, US officials urged observers to ensure that “the performance of the AI ​​system does not fall below an acceptable level over time.”

I’m Bryan Lynn.

Reuters reported this story. Bryan Lynn customized the report for VOA Learning English.


words in this story

artificial intelligence – n. the development of computer systems with the ability to perform work that normally requires human intelligence

to lend – v. Giving someone something for a period of time with the expectation that they will return it later

Also Read :  The Procurement Bill is a chance for the UK to put its money where its mouth is on human rights

fraud n. the crime of using dishonest methods to take something of value from another person; trickery; illusion

Attitude n. a minor change made to something to make it work better

confound v. make someone unable to think clearly or understand something

administer v. to do something or deal with something successfully


What do you think of this story? We want to hear from you. We have a new comment system. Thats how it works:

  1. 1. Write your comment in the box.
  2. 2. Below the box you will see four images for social media accounts. They are for Disqus, Facebook, Twitter and Google.
  3. 3. Click on an image and a box will appear. Enter the login for your social media account. Or you can create one in the Disqus system. It’s the blue circle with the “D” on it. It is free.

Each time you come back to comment on the Learning English site, you can use your account and see your comments and replies to them. Our comment policy is here.


Leave a Reply

Your email address will not be published.