A letter from IBM CEO to Congress has revealed that the company will no longer be offering facial recognition or analysis software to users. The technology that goes into the development of both aspects will also not be researched by the company anymore as it was informed to the public.
The decision of IBM has come from the fact that they oppose the use of any facial recognition technology, which violates basic human rights or the company’s general values. IBM follows principles of trust and transparency, and they believe that now would be the time to decide the fate of facial recognition as delivered by domestic law enforcement agencies.
Artificial intelligence has grown in the past few years, and thanks to that, facial recognition has also seen commendable improvement. Currently, the technology faces a monopoly of privately-owned companies, and there is hardly any federal sight regulating its use. It has often resulted in the technology suffering at the hands of bias created by age, race, and ethnicity. The tools hence prove to be unreliable and unfulfilling of the purpose they were created for.
There have been several pieces of evidence in the past that prove that the bias is spoken of above exists. Research in 2018 showed the extent to which the facial recognition systems were biased, and that included IBM. The algorithms used in the creation of such technology have been criticized highly because of such evidence.
In 2019 as well, a study found the existence of demographic differences in the facial recognition systems that were existing then. Privacy violations have also been a concern among many others when it comes to facial recognition technology.
Many discrepancies have been reported in similar facial recognition systems in the past. Rekognition, which is software sold to law enforcement by Amazon, matched 25 congress members to faces picked from elsewhere. Like this, many companies have come under scrutiny because of the foundation of their facial recognition systems. One of them is Clearview AI, which developed its technology by scraping social media sites, but it now faces numerous privacy lawsuits. Facebook has also been one of the notable companies to have paid $500 million for unlawful use of facial recognition.
However, IBM has tried to reduce the impact that the technology has had by releasing a public data set, which could reduce the bias for a facial recognition model in the future. It was in 2018, but in 2019, the company also released one million photos that came from Flickr, and no consent was taken from the owners here. The data set they had was justified with the statement that it was all taken from public profiles.
Today, in 2020, IBM has finally decided to let go of this segment of technology while advocating that more police misconduct cases come to light. The steps surely reflect the current situation in America and across the world.
Image source: TheVerge