A new generation of artificial intelligence that provides a "confidence rating" of its answers could make algorithms more trustworthy and speed their rollout in safety-critical situations.
Scientists developing the technology called "uncertainty-aware AI" say it removes the risk of hallucinations - fictitious answers created by computer algorithms when they have incomplete or conflicting data.
The new AI assistant would instead provide its best assessment of a situation when faced with patchy evidence, but confirm that it wasn't sure.
A human operator would then be able to judge how much weight to put on the answer and seek additional information where necessary.
The start-up digiLab, which devised the AI assistant called the Uncertainty Engine, is already collaborating with a major manufacturer in the design of a new generation of small-scale nuclear reactors that are likely to be placed closer to people's houses in future.
It helps engineers accelerate development without taking risks.
Read more about AI:
How most people are using ChatGPT
Meet Diella: Albania's AI-generated 'minister'
Tim Dodwell, company founder, told Sky News: "As we adopt AI in situations that are going to [be] transformative, like energy, maritime safety and healthcare, uncertainty really matters.
"I see no real future for AI unless we answer another problem: can we trust it? Human lives are at stake and people have to take responsibility for what the algorithms will do."
Sky News was invited to join a team testing the Uncertainty Engine out at sea, where they believe it could be a game-changer for complex maritime route planning.
The AI assistant is being used on a high-performance sailing yacht to help plot the fastest or safest route.
It can't currently be used in races, but it is helping the Next Step Racing team prepare and understand how to extract even more power from the boat.
Joss Creswell, co-skipper, said offshore racing, like Formula One, is all about small margins - and AI can help crunch a mass of data.
"There's a lot, a lot of information coming to the screen," he said.
"You've got information about the wind, the sea state, the boat's performance, as well as the position of other boats you're racing against.
"Quantifying that and reducing it to something that we understand and then react to is really difficult, so it's really useful to use AI for that.
"It can give you insights into the changes you can make to give you an advantage over the competition."
Read more from Thomas Moore:
AI model could predict disease risk a decade before symptoms
Could refreezing the Arctic actually work?
Mars 'leopard spots' clearest sign life once existed there
Mr Dodwell said the AI assistant could also reduce fuel costs and carbon dioxide emissions of enormous oil tankers and cargo vessels by giving the crew advice on the most efficient route.
"Over a long journey, like Australia to Southampton, there's uncertainty about what the wind and sea conditions will be.
"What we really want is for AI to look at options for the different routes we can take and capture uncertainty in those weather conditions
"It can also capture uncertainty in shipping lanes - for example, if the Suez Canal was closed.
"By doing this intelligently, a 1-5% saving could mean huge benefits to efficiency overall."
Mr Dodwell added that saving 5% in fuel consumption would reduce carbon emissions from the UK shipping industry by 600,000 tonnes a year, equivalent to taking 300,000 cars off the road.
(c) Sky News 2025: New 'AI you can trust' for when safety matters