Two of the world’s leading AI labs, OpenAI and Anthropic, briefly put rivalry aside to conduct joint safety testing of their advanced AI models — a move seen as a rare instance of cross-lab collaboration in an industry defined by secrecy and cutthroat competition.

The project, unveiled on Wednesday, allowed each company’s researchers special API access to versions of their competitor’s models with fewer safeguards, enabling them to test for weaknesses that internal teams might have missed. While OpenAI noted that GPT-5 was not included in the experiment since it had not yet been released, the tests focused on recently deployed models from both companies.

A consequential stage for AI

In an interview with TechCrunch, OpenAI co-founder Wojciech Zaremba said the collaboration reflects the urgent need for industry-wide standards in safety.

“There’s a broader question of how the industry sets a standard for safety and collaboration, despite the billions of dollars invested, as well as the war for talent, users, and the best products,” he said.

He described AI as entering a “consequential” stage of development, where systems are not just research prototypes but products used by millions daily, raising the stakes for safety and alignment.

Nicholas Carlini, a safety researcher at Anthropic, also expressed optimism about the experiment. “We want to increase collaboration wherever it’s possible across the safety frontier, and try to make this something that happens more regularly,” he said.

Competition remains fierce

The cooperation comes against the backdrop of escalating competition between leading labs, where billion-dollar data center investments and $100 million pay packages for top AI researchers have become standard. Experts worry this arms race could incentivize companies to cut corners on safety in order to ship more powerful systems faster.

Indeed, the collaboration did not erase underlying tensions. Shortly after the joint research concluded, Anthropic revoked API access granted to another OpenAI team, accusing OpenAI of violating terms of service by allegedly using Claude to improve competing products. Zaremba insists the incidents were unrelated, but acknowledged that rivalry will remain intense even if safety teams collaborate occasionally.

Key findings: hallucinations and refusals

The research compared how the models behaved in situations where they lacked reliable answers. Anthropic’s Claude Opus 4 and Sonnet 4 frequently refused to answer, declining up to 70% of uncertain questions with responses like, “I don’t have reliable information.”

OpenAI’s o3 and o4-mini models, by contrast, refused questions less often but hallucinated more, offering confident answers even when lacking sufficient knowledge.

Zaremba said the right approach lies between the two extremes — OpenAI’s models should refuse more, while Anthropic’s could attempt to engage more often.

The problem of sycophancy

Both labs also tested for “sycophancy” — the tendency of AI models to agree with users, even when reinforcing harmful behavior. Anthropic’s report flagged examples of “extreme” sycophancy in both GPT-4.1 and Claude Opus 4, where the models initially resisted but later validated concerning or manic user statements. Other models showed lower levels of this behavior.

This issue has recently taken on tragic real-world consequences. On Tuesday, parents of 16-year-old Adam Raine filed a lawsuit against OpenAI, claiming their son relied on ChatGPT, powered by GPT-4o, for advice during a mental health crisis. The chatbot allegedly reinforced suicidal thoughts rather than pushing back, which they say contributed to his death.

Zaremba called the case heartbreaking: “It would be a sad story if we build AI that solves all these complex PhD-level problems, invents new science, and at the same time, we have people with mental health problems as a consequence of interacting with it. This is a dystopian future that I’m not excited about,” he said.

In response, OpenAI said in a blog post that GPT-5 has made significant improvements in reducing sycophancy compared to GPT-4o, particularly in handling mental health emergencies.

Both Zaremba and Carlini say they would like to extend this model of collaboration, testing not just hallucinations and sycophancy but also other pressing safety issues across future AI models. They also expressed hope that other AI developers will follow suit, creating a broader culture of cooperative oversight even as market competition intensifies.

The experiment may be brief, but it highlights a growing recognition among AI leaders that as the technology becomes deeply embedded in daily life, no single lab can guarantee safety alone.

Google has eliminated more than one-third of its managers overseeing small teams, part of the company’s sweeping effort to streamline operations and reduce bureaucracy.

“Right now, we have 35% fewer managers, with fewer direct reports than at this time a year ago,” said Brian Welle, Google’s vice president of people analytics and performance, during an all-hands meeting last week, according to audio obtained by CNBC. “So a lot of fast progress there.”

The cuts affect managers who oversaw fewer than three employees, according to a person familiar with the matter. Many of those affected have remained at the company as individual contributors. Executives said the restructuring is aimed at speeding up decision-making and ensuring managers, directors, and vice presidents make up a smaller share of Google’s total workforce over time.

Job security questions linger

At the meeting, employees pressed Welle and other executives on issues ranging from job stability to workplace culture, following several rounds of layoffs, buyouts, and reorganization efforts over the past year.

Google CEO Sundar Pichai acknowledged the concerns but noted the need for efficiency. “We need to be more efficient as we scale up so we don’t solve everything with headcount,” he said.

Alphabet, Google’s parent company, laid off about 6% of its global workforce in 2023 and has since made targeted cuts across multiple divisions. Chief financial officer Anat Ashkenazi, who joined last year, said in October she planned to push cost-cutting measures “a little further.”

Buyouts gain traction

As part of the restructuring, Google has offered “Voluntary Exit Program” (VEP) packages in 10 product areas, including search, marketing, hardware, and people operations. Between 3% and 5% of employees in those units have accepted the offers, executives said.

“This has been actually quite successful,” said chief people officer Fiona Cicconi, noting that many employees who took buyouts were seeking career breaks or time to care for family members.

Pichai defended the move, saying the buyouts reflected employee feedback. “It gives people agency, and I’m glad to see it’s worked out well,” he said.

Sabbatical debate

The town hall also touched on employee benefits. Workers asked whether Google might adopt a sabbatical policy similar to Meta’s “recharge” program, which grants a month off after five years at the company.

Alexandra Maddison, Google’s senior director of benefits, dismissed the idea, saying the company’s current vacation and leave policies already provide sufficient time for rest.

“We’re very confident that our current offering is competitive,” she said.

Cicconi added that, unlike Meta, Google offers the voluntary buyout program. Pichai quipped in response, “Should we incorporate all policies of Meta while we’re at it? Or should we only pick and choose the few policies we like? Maybe I should try running the company with all of Meta’s policies. No, probably not.”

Balancing morale and profits

The restructuring comes at a time when Alphabet’s stock has continued to climb, rising 10% this year after gains of 36% in 2024 and 58% in 2023. Yet employees say morale has been strained, as the company posts record earnings while continuing to cut jobs.

Executives maintain that the changes are necessary to ensure Google remains agile in a highly competitive tech industry, even as the company grapples with how to balance efficiency with employee trust.

Nigeria’s economic engagements with China have taken on a new dimension, with trade between Nigeria and China’s southern Guangxi region reaching $320 million in 2024, underscoring Beijing’s strategy of leveraging its provincial hubs to deepen African ties.

The disclosure came from Zhang Xiaoqin, Vice Chairman of the Standing Committee of the Guangxi People’s Congress, during a visit to the Nigeria-China Strategic Partnership (NCSP) in Abuja. Zhang led a high-level delegation, joined by Zhou Hongyou, Head of Mission of the Chinese Embassy in Nigeria, in what observers see as part of a broader push by China to embed its subnational governments more firmly into Africa’s economic corridors.

Guangxi’s Rising Profile

According to Zhang, Guangxi is positioning itself as a strategic gateway in southern China, hosting the China-ASEAN Expo and serving as a growing trade hub. He noted that trade with Nigeria rose 21% last year, cementing Guangxi’s role in Beijing’s “regional diplomacy” approach.

The Vice Chairman extended invitations to Nigerian businesses and tourists, citing Guangxi’s strengths in fruit production, mining, digital economy, and tourism—sectors China is keen to internationalize through partnerships. The region, long overshadowed by industrial powerhouses like Guangdong and Jiangsu, is now being pitched as a complementary partner for African economies looking to diversify beyond oil and raw materials.

Mr. Zhou Hongyou of the Chinese Embassy emphasized the cultural and economic similarities between Guangxi and Nigeria, stressing that subnational-to-national linkages could accelerate trade diversification. His remarks mirror Beijing’s strategy of embedding “people-to-people” narratives into economic diplomacy—a tactic often deployed to win African goodwill.

Nigeria’s Strategic Calculus

On the Nigerian side, Joseph Olasunkanmi Tegbe, Director-General of the NCSP, used the visit to spotlight Nigeria’s upgraded ties with China, now elevated to a Comprehensive Strategic Partnership following the 2024 Forum on China-Africa Cooperation (FOCAC).

Tegbe reassured the delegation that the NCSP would continue to monitor agreed projects under FOCAC, oversee strategic investments of national interest, and coordinate Presidential initiatives in manufacturing, agriculture, and infrastructure. He pledged to lead an NCSP mission to Guangxi to “learn from China’s model of large-scale modernization,” framing the visit as part of a broader knowledge-exchange agenda.

Crucially, Tegbe positioned Guangxi as a new partner alongside Hunan Province, pledging to pursue practical opportunities in agricultural processing, manufacturing equipment, and fruit value chains—areas Nigeria is desperate to develop in its industrialization push.

Where the $320m fits in the map of China–Africa trade

Viewed against the aggregate of China–Africa commerce, the Guangxi–Nigeria number is a sliver. China’s trade with the African continent runs in the tens of billions of dollars; provincial figures such as Guangxi’s bilateral tally are therefore best read as complementary slices of a much larger whole.

Guangxi’s engagement is notable not because it reorders the ledger of China’s biggest partners, but because it signals two practical shifts: first, greater provincial initiative in Africa beyond the coastal manufacturing hubs; second, a diversification of China–Africa trade that now includes more regional Chinese players and more targeted, sectoral projects.

This latest engagement comes against the backdrop of $20 billion in investment commitments that Tegbe previously announced in July 2025, secured through recent Nigeria–China negotiations. At the time, he stressed that strengthening bilateral collaboration with Beijing was central to driving Nigeria’s industrialization agenda and reducing reliance on volatile oil exports.

The NCSP has increasingly become the anchor of Nigeria’s China strategy, tasked not only with implementing FOCAC projects but also with forging parallel partnerships outside formal frameworks. Analysts say this dual-track approach is necessary, given that many FOCAC projects often face implementation delays.

A Shift in China’s Africa Playbook

What is striking about Guangxi’s engagement is how China is decentralizing its Africa strategy. While Beijing and major coastal provinces like Guangdong have traditionally dominated Africa’s trade relations, provinces like Guangxi and Hunan are now carving out roles of their own. This allows China to expand Africa engagement beyond megaprojects into more localized, sector-driven collaborations.

For Nigeria, this diversification could bring advantages—Guangxi’s focus on fruit and agriculture aligns with Abuja’s need to strengthen food security, while its mining experience dovetails with Nigeria’s push to revive its solid minerals sector.