TikTok Ban Might Be Attached to House Ukraine and Israel Aide Package

Report from New York Times

In Brief – Speaker of the House Mike Johnson (R-LA) is reportedly trying to include legislation to force a sale of the US TikTok business by its China-based parent company ByteDance in a legislative package providing military aid to Ukraine and Israel. While the Senate passed a combined Ukraine and Israel aid bill two months ago, the package has been stalled in the House due primarily to opposition to the Ukraine assistance by a sizeable number of conservative Republicans, as well as opposition to military assistance to Israel by some progressive Democrats. Along with Ukraine and Israel aid, the still fluid legislative package may also include humanitarian assistance to Palestinians in Gaza, support for Taiwan, measures to address the US southern border, and a version of the TikTok bill passed in March. While the initial TikTok bill gave ByteDance six months to complete a sale, the new version will allow for up to 12 months. Key Senate Committee Chair Maria Cantwell (D-WA), who recently called for extending the TikTok timeline to 12 months, said she supported the modified measure. While the House is expected to vote on the final components separately, the procedural framework will likely automatically combine the package and present them to the Senate as a single measure. President Biden urged quick passage of the aid package.

Context – The initial House TikTok vote showed some opposition from both Republicans and Democrats. And there were clearly some concerns in the Senate as well. Running out the clock in that less-transparent legislative body is a tried-and-true strategy. However, if a TikTok ban ends up as part of a Ukraine-Israel package that actually passes the House, it will likely circumvent any Senate hurdles. But Congress was never the main firewall for TikTok or its users. The 1st Amendment and federal courts are. When the Trump Administration tried to ban TikTok and WeChat in 2020, three federal judges imposed temporary halts primarily on 1st Amendment ground. Most noteworthy, a judge ruled that Americans had the right to use WeChat and that platform is clearly a part of China’s digital influence and censorship regime. TikTok says they are not.

Meta Oversight Board to Consider Policies on AI-Created “Deepfake” Nudity

Report from the Reuters

In Brief – The quasi-independent Meta Oversight Board, made up of noted outside experts from around the world, has agreed to review how the company handles explicit, AI-generated images. The Board announced that it is looking into two specific instances, one involving Instagram in India, and the second Facebook in the US, where the company was confronted with AI-generated images of female public figures shown in varying degrees of nudity. In both cases, users asked that the images be taken down, and eventually both were. However, the Board says that the image of the American woman was taken down by Facebook much quicker than the image made to look like a well-known Indian woman was taken down by Instagram in that market. Meta says that it uses a mix of artificial intelligence and human reviewers to detect sexually suggestive content. The Board is taking public comments until April 30, including on the harms of deepfake porn, the proliferation in different regions, and weaknesses of Meta’s approach to deal with it.

Context – Hundreds of millions of people post content on digital platforms every day… endless video, audio, images, and text. The platforms all have rules, some more strict, some less strict, trying to create their version of a successful platform. While debates about rules and content moderation decisions often focus on a few politically salient topics, the number of nuances, grey areas, and mind-boggling scale issues, never ends. Super smart tech policy expert Mark Masnick describes the content moderation job as “Impossible to do well”. The Meta Oversight Board is the most innovative attempt to improve the process without resorting to government orders. Ironically, while the panel is delving into the increasingly high-profile issue of “deepfake” nudes, Meta has historically been more aggressive restricting nudity and sexually explicit content than most big social media platforms. In fact, if you scan past Oversight Board decisions, you’ll see that they’ve taken on a number of cases involving policies on nudity, generally encouraging Meta to be more flexibility, nuanced, and better consider context.

Google Exploring Hiding News Story in California to Address News Link Tax

Report from AP

In Brief – Google has announced that it is testing a modified version of Google search on a small share of users in California that does not deliver links to the kind of news stories that would require Google to pay news media companies if the state enacted the California Journalism Preservation Act (CJPA). The CJPA forces large digital platforms such as Facebook and Google to pay news media companies a “journalism usage fee”, which Google calls a “link tax”, when news content they publish appears on the platforms next to advertising. The measure was passed by the State Assembly last year. The State Senate has until the end of 2024 to pass the measure before the legislation expires and the process would need to restart in 2025. Meta has said they would block covered news links in California if the law were enacted.

Context – A confrontation over legislation that forces Google, Meta, and other giant digital platforms to pay media companies when users post, share, or access news content has been building in the US for years. The divergence between Meta’s and Google’s policies, which has manifested in Canada, is the central drama. Backing up, Australia enacted a forced media payments bill in 2021. Meta and Google threatened to block news content and Meta did for a few days. The Australian legislation was modified, both companies signed agreements to pay millions of Australian media companies, and the conflict settled down. Canada enacted similar legislation last year. Again, both companies had threatened to cut off news links. Meta did. Their blocks remain in place. Traffic to their top platforms did not decline, but media sites have suffered traffic losses. Google instead agreed to pay $100 million Canadian annually into a news fund. Back in Australia, the platforms’ media deals need to be renewed. Meta is expected to stop payments, implement a Canada-style policy, and set off a row. If California, or the US Congress, were to enact a news company payment regime, Meta’s course seems set. Google needs to consider whether they want to be the sole payer, or whether a Meta-like policy works for them. Is TikTok willing to pay?

UK CMA Outlines Latest Competition Concerns with AI Development

Report from the TechCrunch

In Brief – The UK Competition and Markets Authority (CMA) believes that investments by the largest digital companies in AI startups is raising the prospect that the emerging industry could become dominated by the same digital giants that exercise immense market power in the internet economy. A new report by the CMA on the development of AI foundation models, which follows on an initial report released last fall, says that an “interconnected web” of partnerships and investments within AI involves a handful of the largest tech firms that may be able “to shape these markets in their own interests.” Microsoft, Meta, Google, Apple, Amazon, and AI-chip giant Nvidia are highlighted, including Microsoft’s massive backing of OpenAI and French leader Mistral, and both Google and Amazon investments in Anthropic, a top OpenAI challenger. Chipmaker Nvidia has backed dozens of startups including LLM builder Cohere.

Context – One mantra of tech regulation advocates is that governments should not make the “same mistakes” with AI that they made with social media, apparently meaning regulating more and sooner. On competition policy, they often say AI is being dominated by the same digital giants. However, startup OpenAI is the leader in LLM chatbots, having outperformed the homegrown models of the internet giants, while Clearview AI pulled off a similar trick beating the giants on facial recognition. Regardless, the EU Competition Authority, UK CMA, and US FTC are all on the AI beat, in particular studying Microsoft’s relationship with OpenAI. The CMA’s reticence to trust the market on AI is another example of the UK’s schizophrenic effort to be a pro-innovation alternative to the EU while pursuing their own flavor of aggressive regulatory interventions. The proposed UK Digital Markets Competition and Consumer Bill, a 2024 legislative priority, parallels the EU Digital Markets Act. The UK’s Online Safety Bill (OSB), passed last year, is the UK take on the EU Digital Services Act. And how long can the UK Government press forward with dire AI safety warnings while claiming they won’t regulate AI "in the short term".