W
Welcome to TechCrunch Exchange, a weekly newsletter about startups and markets. Inspired by his daily TechCrunch+ column from which it takes its name. Want it delivered to your inbox every Saturday? Sign up here.
Ability to reinvent technology Wheels have their drawbacks. It can mean ignoring obvious truths that others already know. But the good news is that a new founder may figure it out on his own faster than his predecessor. — anna
AI, trust and safety
This year is an Olympic year and a leap year. . .and again of election year. But before you criticize me for US debt defaultism, I'm not just thinking of a Biden vs. Trump sequel. National elections are being held in over 60 countries, not to mention the EU Parliament.
Changes in each of these votes could impact technology companies. For example, different political parties tend to have different views on AI regulation. But even before an election is held, technology will play a role in ensuring election integrity.
Election integrity may not have been in Mark Zuckerberg's mind when he founded Facebook and perhaps even when he bought WhatsApp. But 20 years later and his 10 years on, respectively, trust and safety have now become inescapable responsibilities for Mehta and other tech giants, whether they like it or not. This means working to prevent disinformation, fraud, hate speech, CSAM, self-harm, and more.
But AI could make this task even more difficult, and not just because of deepfakes and empowering a large number of bad actors. Lotan Levkowitz, general partner at Grove Ventures, said:
All these reliable and secure platforms have hash sharing databases, so you can upload your bad stuff there and share it with all the community. That way, we can all work together to stop it. But today we can train models to avoid that. So even the more classic trust and safety efforts are becoming increasingly difficult thanks to Gen AI, as algorithms can bypass all of this.
From an idea to the forefront
Online forums have already learned a thing or two about content moderation, but when Facebook was born, there was no social network strategy to follow. Therefore, it is somewhat understandable that it will take some time to tackle this challenge. However, it is disheartening to learn from internal meta-documents that as of 2017 there was still an internal reluctance to introduce measures that could better protect children.
Mr. Zuckerberg was one of five social media technology CEOs who recently attended a U.S. Senate hearing on children's online safety. Testimonials are not new to Meta so far, but the inclusion of Discord is also notable. Although it has expanded beyond its gaming roots, it serves as a reminder that trust and safety threats can occur in many places online. This means that in social gaming apps, for example, users may be at risk of phishing and grooming.
Will new companies acquire stock faster than FAANG? That's not guaranteed. Founders often act on first principles, which can be a good thing or a bad thing. The learning curve for content moderation is real. But OpenAI is much younger than Meta, so it's reassuring to hear that it's forming a new team to study child safety, even if it's the result of increased scrutiny.
But some startups wait for signs of trouble and don't take action. His ActiveFence, a provider of AI-powered reliability and safety solutions and part of the Grove Ventures portfolio, is seeing an increase in inbound requests, his CEO Noam Schwartz told me. Told.
“We've seen a lot of people approach our team from companies that are just established or haven't even launched yet. They're thinking about product safety at the design stage. [and] Adopts the concept of safety by design. They are building safety measures within their products, just as they think about security and privacy when building features today. ”
ActiveFence isn't the only startup in what Wired describes as “trust and safety as a service.” But because the company is one of the largest, especially since its acquisition of Spectrum Labs in September, its clients include large companies that fear a PR crisis or political scrutiny, as well as small businesses just starting up. Good to hear that the team is included as well. The technology industry also has an opportunity to learn from past mistakes.