🌟OpenTools Special Edition

AI companies bypassing web protocol to scrape publisher content, claims licensing firm

Key Points 

• According to content licensing startup TollBit, AI agents, from various companies, are ignoring a standard web protocol that publishers use to prevent their content from being scraped for generative AI systems.

• This pattern becomes more apparent with increased analysis of publisher logs.

☕News - According to content licensing startup TollBit, several AI companies are bypassing a standard web protocol that publishers use to prevent their content from being scraped for generative AI systems. 

In a letter to publishers, TollBit highlighted this issue without naming the specific AI companies or affected publishers. Notably, this disclosure comes in the context of a public dispute between AI search startup Perplexity and Forbes regarding the same web standard, amidst a broader discussion on the worth of content in the era of generative AI.

🙀Highlights - 

• TollBit's letter states that Perplexity is not the only company ignoring the Robots Exclusion Protocol (robots.txt), which dictates which parts of a website can be crawled.

• The letter emphasizes that multiple AI agents, from various companies, are ignoring the robots.txt rules to scrape content from websites, and this pattern becomes more apparent with increased analysis of publisher logs.

• The News Media Alliance, representing over 2,200 U.S. publishers, voiced concerns about the consequences of ignoring "do not crawl" signals. President Danielle Coffey stated that without the ability to block extensive scraping, publishers cannot monetize their valuable content or pay journalists, which could seriously harm the industry.

🥸For context - TollBit is an early-stage startup that aims to connect AI companies seeking content with publishers willing to license it. The company monitors AI traffic on publishers' websites and uses analytics to help negotiate usage fees. 

TikTok accidentally releases AI tool that lets users create controversial content on command

Key Points 

• TikTok accidentally released an internal version of its digital avatar too. 

• CNN was the first to notice this issue and used it to create videos featuring quotes from Hitler and a message encouraging people to drink bleach.

• Importantly, none of these videos included a watermark indicating they were AI-generated. 

😵‍💫News - TikTok accidentally shared a link to an early version of its AI digital avatar tool without restrictions, allowing users to make videos containing controversial content.

CNN was the first to notice this issue and used it to create videos featuring quotes from Hitler and a message encouraging people to drink bleach. 

🤫The lowdown:  Introduced earlier this week, TikTok's Symphony Digital Avatars allow businesses to create advertisements using virtual representations of paid actors. This feature includes AI-powered dubbing, which lets advertisers input scripts for the avatars to say within TikTok's content guidelines. 

Despite being intended for TikTok Ads Manager users, CNN discovered a version that allowed anyone with a personal account to try out the tool. Using this internal tool, CNN generated videos reciting Osama bin Laden’s “Letter to America,” a white supremacy slogan, and misinformation encouraging people to vote on the wrong day. Importantly, none of these videos included a watermark indicating they were AI-generated, unlike the official version of TikTok’s Symphony Digital Avatars.

🫠In conclusion - Even though TikTok has taken down this version of its tool, it raises questions about whether users might find other ways to misuse the digital avatar creator and whether TikTok is prepared to handle such challenges.