Microsoft faces complex technical challenges in TikTok carveout

Short Url
Updated 11 August 2020
Follow

Microsoft faces complex technical challenges in TikTok carveout

  • TikTok is functionally and technically similar to ByteDance-owned Douyin, which is available only in China

NEW YORK: Microsoft Corp’s bid to carve out parts of TikTok from its Chinese owner ByteDance will be a technically complex endeavor that could test the patience of President Donald Trump’s administration, according to sources familiar with the setup.

Trump has given Microsoft until Sept. 15 to put together a blueprint for an acquisition that safeguards the personal data of Americans stored on the short-video app, and he has issued an order to ban it if there is no deal by then.

Microsoft is negotiating a transition period that will give it time to ringfence TikTok technologically from ByteDance after they agree to a deal, Reuters reported on Aug. 2.

The clean break that Trump and lawmakers envision could take a year or more, some of the sources said.

TikTok is functionally and technically similar to ByteDance-owned Douyin, which is available only in China, and shares technical resources with it and other ByteDance-owned properties, people familiar with the matter said. ByteDance started working on their technological separation several months ago amid scrutiny from the US government, a source familiar with the process told Reuters. It began planning for a split as part of a strategy to shift its power from China, Reuters has reported.

While the code for the app, which determines the look and feel of TikTok, has been separated from Douyin, the server code is still partially shared across other ByteDance products, the source said. The server code provides basic functionality of the apps such as data storage, algorithms for moderating and recommending content and the management of user profiles.

To ensure uninterrupted TikTok service, Microsoft would likely need to rely on ByteDance’s code while it reviews and revises the code, and moves to a new back-end infrastructure to serve users, according to cybersecurity expert Ryan Speers at River Loop Security, which provides services including cybersecurity due diligence for deals.

Any continuing technical or operational reliance of the US business on the Chinese company after the sale generally would have been unacceptable to the Committee on Foreign Investment in the United States (CFIUS), said Aimen Mir, former deputy assistant secretary of the Treasury responsible for CFIUS, now a partner at the law firm Freshfields Bruckhaus Deringer.

In the past, CFIUS has required adoption of increased protections pending a sale, including separation of the US business from foreign sellers to the furthest extent possible, he said.

Another challenge Microsoft faces is how it will transfer what is viewed as TikTok’s secret sauce, the recommendation engine that keeps users glued to their screens. This engine, or algorithm, powers TikTok’s “For You” page, which recommends the next video to watch based on an analysis of user behavior.

TikTok uses recommendation algorithms that are independent from Douyin, according to two sources familiar with the matter. But what makes it tick is the content and user information that is fed into the algorithm.

“Algorithms are not worth anything without the data,” said Jim DuBois, a former chief information officer at Microsoft. DuBois is a venture adviser at Ignition Partners. “Segmenting the data for those countries is a significant task.”


WEF report spotlights real-world AI adoption across industries

Updated 19 January 2026
Follow

WEF report spotlights real-world AI adoption across industries

DUBAI: A new report by the World Economic Forum, released Monday, highlights companies across more than 30 countries and 20 industries that are using artificial intelligence to deliver real-world impact.

Developed in partnership with Accenture, “Proof over Promise: Insights on Real-World AI Adoption from 2025 MINDS Organizations” draws on insights from two cohorts of MINDS (Meaningful, Intelligent, Novel, Deployable Solutions), a WEF initiative focused on AI solutions that have moved beyond pilot phases to deliver measurable performance gains.

As part of its AI Global Alliance, the WEF launched the MINDS program in 2025, announcing its first cohort that year and a second cohort this week. Cohorts are selected through an evaluation process led by the WEF’s Impact Council — an independent group of experts — with applications open to public- and private-sector organizations across industries.

The report found a widening gap between organizations that have successfully scaled AI and those still struggling, while underscoring how this divide can be bridged through real-world case studies.

Based on these case studies and interviews with selected MINDS organizations, the report identified five key insights distinguishing successful AI adopters from others.

It found that leading organizations are moving away from isolated, tactical uses of AI and instead embedding it as a strategic, enterprise-wide capability.

The second insight centers on people, with AI increasingly designed to complement human expertise through closer collaboration, rather than replace it.

The other insights focus on the systems needed to scale AI effectively, including strengthening data foundations and strategic data sources, as well as moving away from fragmented technologies toward unified AI platforms.

Lastly, the report underscores the need for responsible AI, with organizations strengthening governance, safeguards and human oversight as automated decision-making becomes more widespread.

Stephan Mergenthaler, managing director and chief technology officer at the WEF, said: “AI offers extraordinary potential, yet many organizations remain unsure about how to realize it.

“The selected use cases show what is possible when ambition is translated into operational transformation and our new report provides a practical guide to help others follow the path these leaders have set.”

Among the examples cited in the report is a pilot led by the Saudi Ministry of Health in partnership with AmplifAI, which used AI-enabled thermal imaging to support early detection of diabetic foot conditions.

The initiative reduced clinician time by up to 90 percent, cut treatment costs by as much as 80 percent, and delivered a 10 time increase in screening capacity. Following clinical trials, the solution has been approved by regulatory authorities in Saudi Arabia, the UAE and Bahrain.

The report also points to work by Fujitsu, which deployed AI across its supply chain to improve inventory management. The rollout helped cut inventory-related costs by $15 million, reduce excess stock by $20 million and halve operational headcount.

In India, Tech Mahindra scaled multilingual large language models capable of handling 3.8 million monthly queries with 92 percent accuracy, enabling more inclusive access to digital services across markets in the Global South.

“Trusted, advanced AI can transform businesses, but it requires organizing data and processes to achieve the best of technology and — this is key — it also requires human ingenuity to maximize returns on AI investments,” said Manish Sharma, chief strategy and services officer at Accenture.