Paul Sweeting

Negotiations toward a final text of the European Union’s AI Act are going down to the wire this week as the final “trilogue” session among the EU Parliament, Commission and Council is scheduled for Wednesday (Dec. 6). The pressure is on to reach an agreement before the end of the year, as the June 2024 EU Parliamentary elections loom over talks. If agreement can’t be reached before then, there’s a danger that the process would have to be restarted with a new Parliament and new leadership in the Council, which could potentially scuttle the whole project.

Yet despite the pressure, the parties to the current talks appear to be farther apart than where they started, endangering what had been touted as the world’s first comprehensive regulatory regime for of AI. The consensus on the basic structure of the proposed regulations that seemed at hand in the summer was thrown into turmoil last month when France, supported by Germany and Italy, suddenly reversed its position and embraced “mandatory self-regulation” via codes of conduct for the largest foundation models instead of the once-agreed tiered system of binding obligations.

Read More »

The distinguishing characteristic of large language models (LLMs) is, as the name implies, their sheer size. Meta’s LLaMA-2 and OpenAI GPT-4 are each comprised of well more than 100 billion parameters — the individual weights and variables they derive from their training data and use to process prompt inputs. Scale is also the defining characteristic of the training process LLM’s undergo. The datasets they ingest and are almost incomprehensively large — equivalent to the entire World Wide Web — and require immense amounts of computing capacity and energy to analyze.

Read More »

Lenin famously said, “There are decades where nothing happens; and weeks where decades happen.” Last week offered a fair approximation of the latter in the world of AI. On Monday, Ed Newton-Rex, the widely respected head of StabilityAI’s audio team, abruptly resigned over a disagreement with the company’s position on the use of copyrighted works to train generative AI models. On Friday, OpenAI even more abruptly sacked its high-profile co-founder and CEO Sam Altman and set a new standard for mangling internal communication and investor relations. And somewhere along the way, Meta reportedly dissolved its Responsible AI team and reassigned its staffers to other units.

The turmoil continued through the weekend. OpenAI president Greg Brockman quit in solidarity with Altman, employees threatened mass resignations and its blindsided investors demanded a restoration and lined up to finance any new AI venture Altman and Brockman might launch. By Saturday evening, talks reportedly were underway between Altman and the company to bring him and his team back into the fold.

Read More »

If you’re wondering where a robust regulatory response to generative AI could come from, don’t sleep on the Federal Trade Commission. While courts and the U.S. Copyright Office are still feeling their way through the complex maze of technical, legal and policy questions raised by the generative AI technology, the FTC, as a law enforcement agency, has a narrower brief and potent tools it already can bring to bear on many of those questions.

It also has a green light from the White House to use those tools. In President Joe Biden’s recent executive order on AI the FTC was encouraged to take a leading role in the government’s efforts to put guardrails around the development and use of AI systems.

Read More »

For the second time in a week, a federal district court judge in California has sent the plaintiffs in an AI copyright infringement lawsuit back to the drawing, dismissing most of the charges brought against Facebook-parent Meta over its Llama generative AI tool while granting them leave to amend and refile their complaint.

The ruling partially granting Meta’s motion to dismiss came down on Friday, one week after another federal judge dismissed most of the copyright infringement charges brought against StabilityAI and other AI image generators while granting leave for plaintiffs to amend and refile their complaint.

Read More »

It’s no secret that the transition to streaming-based economic models for the music, movie and television industries has been disruptive to those businesses. But the force of that disruption has not hit all parties equally or equitably.

The media companies that own music labels and publishers, and the film and TV studios have faced challenges adapting, but at least some of the problems they face have been of their own making. The music companies were rocked, starting with Naptser in 1999, by technology changes they could have, and arguably should have, seen coming. Streaming has been a lifeline since then but they’re now dealing with slower subscriber growth amid market saturation in the major territories.

Read More »

The artists suing AI image generators for copyright infringement suffered a setback this week when a federal district judge in California dismissed most of the charges they brought against Stability AI, Midjourney and DeviantArt. In a 28-page order, Judge William Orrick called the plaintiffs’ complaint “defective in numerous respects,” and noted that two of the three named artist-plaintiffs, Karla Ortiz and Kelly McKernan, had never registered their work with the U.S. Copyright Office and therefore are precluded from bringing an infringement claim in federal court in the first place.

Read More »

European Union officials made incremental progress in negotiations last week over the final text of the EU AI Act, which could establish a de facto global benchmark for regulating artificial intelligence technology and companies, according to local reports, Known as a trilogue, the negotiations, among the EU Parliament, Commission and Council of Ministers, is the final legislative stage before formal adoption of the framework that will be sent to EU member states for implementation in national laws. A fifth and potentially final round of discussions is set for December, but the trilogue could continue if final agreement is not reached.

Read More »

News media publishers, as a group, made what turned out to be a costly error in the early days of the internet. Excited at the prospect of gaining uncounted new readers beyond their home markets they made their reporting openly available online for free. What they did not, and perhaps could not anticipate in those early days, was the monopolistic potency of network effects that would shape the online ecosystem. The result was to concede publishers’ direct connection with readers to search engines and social media platforms and the rupture of their relationship with advertisers.

Read More »

Amazon-backed Anthropic has joined the AI copyright litigation hit parade, catching a four-count infringement lawsuit from a trio of leading music publishers topped by Universal Music Group. The lawsuit was filed Wednesday in federal district court in Nashville. Concord and ABKCO are also named plaintiffs.

Interestingly, the target of the complaint is Anthropic’s large language model (LLM), Claude, which is not a music generator tool and was not trained with musical compositions per se. Instead, the publishers call out Claude for copying and distributing the lyrics to songs owned or represented by the publishers in both its training input and generative output.

Read More »