AI Pact, OpenAI, and TikTok: AI remains at the heart of the news

The past few weeks have been full of developments in the AI world. On one hand, the introduction of the European AI Pact has brought the AI Act and its implementation steps back into the spotlight. On the other hand, news from two industry giants—OpenAI and TikTok—is making headlines, sparking a mix of wonder, concern, and "bad buzz."

The European AI Pact: A first step toward implementing the AI Act?

An AI pact with no binding commitment

While waiting for the AI Act to officially come into force in the summer of 2026, the EU is attempting to rally companies around a European AI Pact. Voluntary companies commit to three main pillars:

• Establishing a governance strategy within the organization.

• Identifying high-risk AI systems.

• Promoting AI literacy within the company.

Any organization can sign the Pact, and giants like Google, Microsoft, and Cisco have already done so. Although the pact is not coercive, signatory companies publicly demonstrate a proactive approach toward AI regulation in their organizations and can gain a head start on complying with the upcoming AI Act.

The giants skipping the Pact: Meta, Apple, TikTok…

Over 120 companies have signed the European AI Pact, but several major players are notably absent, such as Meta, Apple, Mistral, Hugging Face, and TikTok. What should we make of this?

Two scenarios emerge. The first focuses on the nature of the AI Pact itself, which is seen as somewhat superficial and non-binding. Companies like Meta, Apple, or TikTok are likely already advanced enough in their AI integration and internal frameworks.

The second possibility is that refusing to sign is a message of protest, signaling a lack of consensus regarding the content of the AI Act. Unlike the voluntary Pact, the AI Act will have a real impact on these digital giants. 

This could be seen as an opening for further discussion with legislators—an idea reinforced by the departure of Thierry Breton, the architect of the AI Pact, which some interpret as a sign of potential softening in European regulation.

France carves out a place for AI

Since September, France has established a Secretariat of State which, beyond its traditional responsibility for digital affairs, is now specifically dedicated to Artificial Intelligence. This national first suggests a reinforced valuation of AI and digital technology in France. Furthermore, this Secretariat is now attached to the Ministry of Higher Education and Research, rather than the Ministry of Economy. This decision sends a positive message, symbolizing the French government's investment in innovation and R&D.

OpenAI still in the headlines

OpenAI offsets its deficit with a massive fundraise

Currently the world's third-largest private company (behind SpaceX and ByteDance, TikTok's parent company), OpenAI recently raised $6.6 billion, reaching a total valuation of $157 billion. While these figures seem staggering, this fundraiser will likely not be the last.

Unlike Google or Microsoft, which generate profits through software licenses or advertising, AI is OpenAI's sole source of revenue. This makes it difficult to train models and invest in R&D without third-party services to help finance the operation.

With millions of weekly active users, OpenAI is the most widely used solution today. These numbers highlight not only its success but also the immense infrastructure behind it and the colossal costs required for computing power—not to mention the R&D needed to remain competitive against China, which is investing heavily in LLMs. Even if rumors of a ChatGPT subscription price increase prove true, this extra revenue alone would likely not suffice to cover costs.

Missions and Resignations at OpenAI

While high-profile departures, such as CTO Mira Murati and co-founder Ilya Sutskever, fuel concerns about the company's future, other announcements suggest a brighter path forward. For instance, the hiring of Sebastian Bubeck (former VP of AI at Microsoft, specialized in optimizing LLM efficiency) demonstrates OpenAI's commitment to attracting talent and perhaps working on model compression and quantization. Furthermore, addressing criticism, Sam Altman created a dedicated Safety Committee from which he voluntarily excluded himself—a move intended to reassure the public about the safety of the firm's deployed models.

OpenAI launches two new voice and search services

OpenAI has made significant moves recently with two new services:

• ChatGPT "Voice to Voice": Available in France since October 22, the "Advanced Voice Mode" on the mobile app allows for real-time, zero-latency voice conversations. It features a human-like voice capable of understanding emotions, handling interruptions, and even performing roles or accents.
• GPT Search: Last week, OpenAI unveiled "GPT Search," turning its LLM into a full-fledged search engine. In testing, where Gemini (Google’s generative AI) sometimes struggled to provide information, OpenAI synthesized data with remarkable precision while citing all its sources.

TikTok shocks with 700 layoffs in Malaysia

Moderation through Artificial Intelligence

Another tech giant is making headlines following the announcement of 700 layoffs in Malaysia. TikTok has decided to use AI to manage content moderation. The dilemma of machine replacing man is not new, but it is striking once again. While the social and economic impact of these job cuts is significant, one can understand the human impossibility of moderating the world's fifth-largest social network, which has 1.5 billion active users and 34 million videos posted daily.

Entrusting AI with content moderation seems a logical solution already adopted by Meta and Microsoft in 2022 and 2023. Only AI can handle such a scale, allowing for rapid and efficient moderation of the bulk of the content while leaving the most complex cases to humans. While AI offers tireless efficiency, it still struggles to quickly understand evolving languages, irony, cultural references, or trends. In the event of a dispute over content moderation, AI will likely remain less relevant than a human. The challenge lies in finding the right balance in a hybrid model between AI moderation and human arbitration.

These layoffs also likely echo regulatory pressure from the Malaysian government on Big Tech to implement stricter frameworks regarding cybercrime.

These three stories have been the most discussed in recent weeks. Despite the concerns raised, we must stay alert to the long-term impact of these decisions—be it organizational changes at OpenAI or the integration of AI for moderation at TikTok. In the time remaining before the AI Act becomes reality, we will observe what repercussions, if any, the Pact has on the industry.