Signal believes that artificial intelligence is essentially a "surveillance technology"

Why do so many companies that rely on user data to make money seem to have a soft spot for artificial intelligence? If you ask Signal president Meredith Whitaker, she'll tell you it's simply because "artificial intelligence is a surveillance technology."

Taking the stage at TechCrunchDisrupt2023, Whittaker laid out her view that AI is largely inseparable from the big data and targeting industry long-standing by companies like Google and Meta, less consumer-focused but equally prominent enterprise and defense companies.

"It requires surveillance business models; it's an intensification of the growth in surveillance advertising that we've seen since the late 1990s. I think AI is a way to consolidate and expand surveillance business models," she said. "A Venn diagram is a circle. Using artificial intelligence is also a form of surveillance, right? You know, you walk past a facial recognition camera with a pseudoscientific emotion recognition instrument on it and it generates data about you that, rightly or wrongly, says 'you're happy, you're sad, you have a bad personality, you're a liar, etc.' These are ultimately surveillance systems that are marketed to those who have power over us: our employers, governments, border control, etc., to make judgments and predictions that determine our access to resources and opportunities."

Ironically, she points out, the data that underpins these systems are often organized and annotated by the very workers at whom that data can be targeted (a necessary step in the process of assembling AI datasets).

"At the level of informing the fundamental truth about the data, there's no way to make these systems without human labor -- reinforcement learning with human feedback, which again is just a technology sanitizing unstable human labor," she explained. "Thousands of workers are poorly paid, but overall the cost is high, and there's no other way to create these systems, absolutely no other way. In a way, what we're seeing is a Wizard of Oz phenomenon, where when we pull back the curtain, there's not really that much intelligence out there."

However, not all AI and machine learning systems are equally exploitative. When asked if Signal uses any AI tools or processes in its app or development work, she confirmed that the app has a "small device mockup, but we didn't develop it, we used it off the shelf as part of the face blur feature in our media editing toolset. It's not actually that good... but it can help detect faces in crowd photos and blur them so that when you share those photos on social media, you don't reveal people's private biometric data to Clearview and the like."

"But the thing is like... yes, it's a great use of artificial intelligence, and doesn't it allow us to counteract all the negativity that I'm throwing out on stage?" she added. "Of course, if this is the only market for facial recognition... but let's be clear. The process of developing and deploying facial recognition technology is expensive, and the economic incentives will never make this the only use."