Let’s Take Action Before It’s Too Late

In the past few months, Big Tech has come under intense government scrutiny. Facebook, Twitter, and Google are the outlets most often on the hot seat, facing allegations of censorship, deleting and flagging certain content, disabling links, and locking the accounts of people who post content that doesn’t agree with their unwritten agendas. The U.S. Federal Communications Commission and legislators on both sides of the political aisle have called for a review of Section 230, a provision within the 1996 Communications Decency Act that gives internet companies incredible freedom to moderate the content on their platforms. The same companies have also come under fire for how they collect, share, and use the personal information of their users.

Though not as loud as the calls to regulate the social media platforms, speech technology providers—including Big Tech providers like Apple, Google, Microsoft, IBM, and Amazon—are also being placed under the microscope all over the world for how they collect, use, store, and share customer voice data, as our cover story, “Curbing Speech Data Overreach,” points out.

The article raises a really tough question: Has technology reached a point where it’s nearly impossible to rein in, where legislation will never be able to catch up with the speed of advancing technology and innovation? This is a difficult question, and the easy answer is “Probably.”

I’ve known a few people who have fallen victim to a very sophisticated and well-organized scheme involving deep fakes of voice data. Through these scams, the perpetrators gain access to people’s voice recordings (possibly through illegal phone taps) and repurpose the audio files to stitch together a phone dialogue in which the person claims to have been arrested and needs money for bail. The cloned voice files are so convincing that the people targeted by the scam have fallen for it and, in some cases, sent thousands of dollars to help out their seemingly troubled friends or family members.

I seriously doubt that more government regulation will be able to curb that kind of activity; after all, these criminals are already operating outside of the law.

I am not a huge fan of big government and a lot of tight regulations that in the end do little more than slow progress and cripple technological innovation. Without the government’s involvement, the speech industry has made huge strides in just the past few years. The industry has overcome many of the technological challenges that have held it back for decades. Speech recognition, transcription, captioning, translation, and voice search today are nearly flawless, instantaneous, and effortless. That is no small feat, and I would hate to see the pace of advancement slowed by legislation enacted by government bureaucrats alone who know little or nothing about the speech industry.

At the same time, though, I cannot deny the data security risks of having so much speech data, including past utterances to Alexa devices, biometric voiceprints, and search terms spoken to Siri, stored on some server somewhere. One would expect the tech giants to have the most robust security protocols in place, but not every company that works with speech data has access to the same resources as Amazon, Nuance Communications, Apple, or IBM.

The other huge problem with current legislation is that it is wildly inconsistent. “Regulation of voice data is a patchwork of laws and regulations around the world,” laments Debbie Dahl, Speech Technology’s resident expert on industry standards.

In many cases, the default position is to let the industry regulate itself, which has only been mildly successful so far.

“Because of technology’s rapid advancement, we cannot rely on companies alone to make the ethical decisions,” Nava Shaked, a longtime speech industry insider, says in the article. “The challenge will be creating and passing legislation in a way that will not delay technology from happening or scare users from cooperating.”

It’s a tall order, but I am convinced that it can be done if government, industry, and consumers all work together. Hopefully, that happens before a large-scale voice data breach does irreparable damage and permanently erodes whatever trust consumers might have in the technology and the companies that wield it.

Leonard Klie is the editor of Speech Technology magazine. He can be reached at lklie@infotoday.com.