Internal deliberations reported within Meta Platforms Inc., the parent company behind Facebook and Instagram, reportedly show the company’s aggressive pursuit of digital data to fuel its artificial intelligence advancements. Amid the competitive race to dominate AI technology, strategists at Meta, which include managers, lawyers, and engineers, have apparently explored unconventional avenues, such as the acquisition of prominent publishing house Simon & Schuster.

Read my report for ReadWrite on the NYT probe: OpenAI and Google accused of using YouTube transcripts for AI

According to leaked recordings of internal meetings acquired by The New York Times, the tech giant’s desperation for a rich dataset to power its AI algorithms has led to discussions that possibly straddle ethical boundaries. Ahmad Al-Dahle, Meta’s Vice President of Generative AI, purportedly conveyed to executives that the vast expanses of the internet had been scrubbed to feed the development of a sophisticated model.

The pursuit of data

Al-Dahle reportedly stated that rhe company had already used almost every available English-language book, essay, poem, and news article on the internet, in a bid to compete with rival technologies like ChatGPT.

Read: Authors’ pirated books used to train Generative AI

It is then claimed that the team at Meta considered the purchase of Simon & Schuster, a leading publisher with a roster including household names like Stephen King.

In October 2023, the private equity firm KKR successfully completed its $1.62 billion acquisition of Simon & Schuster. KKR emerged as the winning bidder in early August, after the government rejected Penguin Random House’s attempt to acquire S&S in late 2022.

Ethical and legal boundaries ‘tested’

The discussions also allegedly veered into legally murky territories, with talks of aggregating data from the web potentially at the risk of litigation. The idea that negotiating licenses would take too long was reportedly echoed across several meetings.

This aggressive strategy has not been without its detractors. During one of the meetings, a lawyer is said to have raised “ethical concerns” regarding the appropriation of intellectual property, only to be met with silence.

At the heart of these discussions is Meta’s founder, Mark Zuckerberg, who has been vocal about the imperative for Meta to enhance its AI capabilities. “The capability that Mark is looking for in the product is just something that we currently aren’t able to deliver,” an engineer admitted during the discussions.

Meta’s push into AI comes at a time when the company faces significant data limitations due to a combination of user behaviour and self-imposed privacy protocols. Many users have pruned their historical posts, and the platform itself, traditionally not a repository for lengthy, essay-type content, offers limited fodder for AI development. On top of this, the privacy overhaul following the 2018 Cambridge Analytica scandal has further restricted access to user data.

In pursuit of a workaround, it was reported that Meta employed contractors in Africa to compile summaries of copyrighted materials, a practice that skirts close to legal infringements. “Because we have no way of not collecting that,” a manager explained, showing the challenges of distinguishing between copyrighted and non-copyrighted material in data aggregation efforts.

Response from authors

A number of authors have been alarmed at the report, calling for further copyright protection of their work. New York Times bestselling author Kelly Yang wrote on X, formerly known as Twitter, that these protections should be built into contracts. While writer Ewan Morrison said that Big Tech should be broken up.

The leaked recordings paint a picture of a company at a crossroads, grappling with the ethical and legal implications of its quest for AI supremacy.

How To Be Books has approached Meta for a comment and is awaiting a response.