TOKYO - Japanese companies and venture investors are stepping up efforts to commercialize technologies that could transform shipping and recruitment, with Mitsui O.S.K. Lines pursuing fully autonomous operation of giant cargo vessels while startup Calaris develops an interactive AI interviewer designed to uncover abilities that conventional job applications may miss.
Mitsui O.S.K. Lines, one of Japan's leading shipping companies with a history spanning about 140 years, operates around 900 vessels worldwide, including large cargo ships and cruise ships.
Its corporate venture capital arm MOL PLUS was established in 2021 with a 4 billion yen investment allocation from its parent company and has invested in startups developing technologies that could shape the future of shipping.
The company also opened a satellite office in Tokyo in 2024 designed as a place where employees and outside partners can exchange ideas and discuss innovation.
The scale of MOL's shipping operations is reflected in some of its largest vessels, which stretch more than 300 meters, roughly comparable to the height of Tokyo Tower.
MOL is now combining the activities of two corporate venture capital operations as it considers investment priorities for the next five years, with autonomous vessel operation emerging as one of its key targets.
The second operation, MOL Switch, was established in Silicon Valley in 2023 and has invested about 12 billion yen in overseas startups over three years, focusing particularly on the energy sector.
MOL believes advances in autonomous driving on land could eventually be applied to ships, potentially allowing large commercial vessels to operate with much less direct human control within the next 10 to 20 years.
The technology could also improve crew safety as geopolitical instability creates greater risks for vessels operating in some parts of the world.
However, fully autonomous shipping presents challenges that differ significantly from self-driving cars.
Autonomous vehicles can combine GPS, radar, cameras and other sensors to identify hazards and respond quickly. A large cargo ship, by contrast, can continue traveling for several kilometers through inertia even after its engines are stopped.
Stopping distance also changes depending on waves, water depth, wind, currents and the type and amount of cargo being carried.
MOL has identified startups specializing in individual technologies such as image recognition and automated positioning, but has yet to find a single company that possesses all of the technology required for fully autonomous operation.
That means one of the group's biggest challenges will be determining which startups to invest in and how their separate technologies can ultimately be combined into a functioning autonomous navigation system.
Data collection presents another obstacle. Self-driving cars can repeatedly travel along roads to gather information and train their systems, but conditions at sea are difficult to reproduce.
Waves, swells, wind and currents can be different each time a vessel approaches a port, requiring data to be collected under a wide variety of conditions. Navigation also becomes particularly complicated when ships enter congested coastal waters or areas with large numbers of fishing vessels.
If MOL succeeds in developing the technology, it sees the possibility of eventually providing autonomous-navigation systems to shipping companies around the world, effectively creating a maritime equivalent of an autonomous-driving service such as Waymo.
The company expects the number of crew members physically working aboard ships could decline as automation advances, but says it intends to continue training seafarers while simultaneously preparing for a more autonomous future.
MOL's startup investments are also extending beyond conventional shipping. One project under consideration involves vessels serving as offshore launch and landing platforms for reusable rockets.
Japan has limited land available for the safe operation of reusable rockets that return to Earth, making offshore facilities a potential alternative. MOL envisages buying and owning the vessels and charging space companies for individual launch or landing operations over an operating life of roughly 20 years.
A separate wave of innovation is emerging in recruitment, where generative AI is making it increasingly difficult for companies to distinguish applicants using conventional resumes, entry sheets and written examinations.
At the Tokyo Venture Capital Hub in Azabudai Hills Garden Plaza B, more than 70 venture capital firms have established operations and exchange information while seeking promising startups.
Among them is the University of Tokyo Edge Capital Partners, commonly known as UTEC.
UTEC was established in 2004 when Japan's national universities were incorporated and describes itself as the country's first university-affiliated venture capital firm. It works closely with the University of Tokyo on areas including the commercialization of patents and university-developed technologies, while also investing in companies with no direct connection to the university.
The firm focuses on startups creating new value through science and technology and has invested in more than 160 companies, with its funds totaling more than 130 billion yen.
One of its latest investments is Calaris, a startup founded in 2025 that is developing products using generative AI.
Calaris says advances in generative AI have created a new problem for recruiters because applicants can use AI to polish resumes and application documents to such a degree that differences between candidates become difficult to identify.
The startup has developed an interactive AI assessment tool called Calaris Assess that conducts interviews intended to reveal a candidate's underlying abilities through extended dialogue.
Instead of relying only on standard questions for which applicants can prepare answers in advance, the system begins with familiar topics before asking increasingly detailed follow-up questions based on the candidate's responses.
Calaris says around 30 minutes of conversation is needed for a full assessment because deeper questioning makes it progressively harder for applicants to rely on prepared answers.
The company began full-scale commercial provision of the service in June, and the system has already been introduced by around five or six companies.
Calaris raised 600 million yen in a seed financing round, with UTEC providing the majority of the investment.
After a 30-minute assessment, the system produces an overall score along with evaluations across categories including the ability to think, mobilize other people, carry tasks through to completion and continue learning.
It is also designed to identify candidates with unusually strong abilities in particular areas even when their overall scores are lower.
For example, a person who performs poorly in structured or logical thinking could still receive a particularly high score for creative thinking, potentially allowing employers to identify applicants who might have been rejected under conventional recruitment systems based primarily on overall performance.
Calaris deliberately excludes factors such as a candidate's atmosphere, speaking style or general personal impression from the AI assessment.
The company does not envisage AI replacing human recruiters entirely. Instead, the automated assessment is intended to help identify underlying abilities during the initial selection process, with people conducting second and third interviews to judge factors such as personality and compatibility with the company.
The approach could create a division of labor between AI and human recruiters, using technology to identify abilities that are difficult to see on paper while leaving final hiring decisions and assessments of personal fit to people.
Source: テレ東BIZ
