Latest news with #WilliamTunstallPedoe
Times
11 hours ago
- Business
- Times
Can AI be made trustworthy? Alexa inventor may have the answer
One of the inventors of Amazon's Alexa has proven he can make AI trustworthy — at least when it comes to assessing valid insurance claims. William Tunstall-Pedoe originally developed the technology that became the retail giant's voice assistant service and his new venture, called UnlikelyAI, has an even more ambitious goal. 'We are tackling a problem that is potentially bigger than Alexa, which is making AI trustworthy,' he said. His company has combined data-driven learning models, known as neural networks or large language models (LLMs), with rule-based systems, called symbolic reasoning, to create a platform that companies can use to automate their processes using AI. 'LLMs have amazing capabilities and are absolutely transformative but when enterprises try to apply LLMs to problems in their business it very often doesn't work,' said Tunstall-Pedoe, 56. 'A lot of pilots don't really succeed. It is a black box, isn't explainable, and it is inconsistent. We are developing fundamental technologies to tackle that problem.' UnlikelyAI has completed a pilot with SBS Insurance Services, which saw the insurer automate 40 per cent of its claims handling with 99 per cent accuracy. This compares with a rate of accuracy for the same task that is typically around 52 per cent when just using LLMs, the company said. UnlikelyAI's system also provides an audit trail for all its decisions, so they can be explained if queried by customers or regulators. 'We are building a collection of technologies that bring trust to AI applications. Whenever enterprises are using AI to do business critical things, where the cost of getting it wrong is high, we can help,' said Tunstall-Pedoe. 'In the insurance world we are ingesting the policies, which are natural language. We create a symbolic representation of it, which then gives you that really high accuracy when doing the claims process against it.' He sold the technology that became a key part of Amazon's Alexa voice assistant in 2012. It originated in a startup he founded in Cambridge called True Knowledge, which became known as Evi after it developed a voice assistant, a few months after Apple launched Siri. 'We were competing directly with the biggest company in the world as a 30-person Cambridge startup. We had millions of downloads very quickly and every big company that was trying to figure out its response to the existence of Siri were talking to us. At the end of 2012 we had two acquisition offers and we chose to get bought by Amazon.' Tunstall-Pedoe joined the Amazon team to develop Alexa, working on an initiative under the Project D codename, and launching it in the US in 2014. He left Amazon in 2016 and has since invested in over 100 start-ups and mentored entrepreneurs. He founded UnlikelyAI in 2020 and has since raised $20 million from investors including Amadeus Capital Partners, Octopus Ventures, and Cambridge Innovation Capital. Tunstall-Pedoe said UnlikelyAI's 'goal is to create AI that is always right'. 'When it gives you an answer you can always trust it. It can always provide a fully auditable explanation for any business decision that is made. And it will be consistent, and not breach your trust by giving a different answer each time you use it.' 'Our primary customers are high stakes industries, where a business decision has really big consequences if it's wrong. Medicine is a good example. Finance is also very important, or any industry that is regulated. If you breach regulations you can be fined.'
Entrepreneur
14-06-2025
- Business
- Entrepreneur
Entrepreneur UK's London 100: Unlikely AI
Industry: Artificial Intelligence UnlikelyAI is pioneering a novel approach that fuses large language models with symbolic methods to boost AI accuracy, safety, and transparency. In fields like healthcare, it's building technology that delivers explainable, verifiable insights - maximizing AI's benefits without compromising safety. Founder William Tunstall-Pedoe previously created Evi, the tech behind Amazon Alexa.
Yahoo
14-06-2025
- Business
- Yahoo
Meet the Briton who helped teach Alexa how to talk
British AI entrepreneur William Tunstall-Pedoe has always been 'motivated by impact'. He started pushing the boundaries on what's possible with machines as a 13-year-old schoolboy, while Tunstall-Pedoe secretly helped create and launch Amazon's (AMZN) Alexa over a decade ago. In between, his anagram technology was used by author Dan Brown for the Da Vinci Code and he even identified the most boring date in history. The first keyboard seeds were sown at Dundee High School in Scotland where he was able to venture to the technical college next door before classes and during breaks to write computer software on its mainframe. There were financial rewards, too. His computer science teacher, Michael Ryan, had a software business which sold by mail order. However there was no taking advantage of the pupils, says Tunstall-Pedoe, with very generous royalties and a decent income on offer. Read More: Meet Britain's 'king of billboards' who sold his business for £1bn 'I kind of very arrogantly set out to try and solve this problem of using language to control machines,' he tells Yahoo Finance UK, 'to automatically answer questions, change the user experience you have when you interact with search engines, when you interact with computer software, and kind of invented some technology to tackle that.' It is 20 years since he started his search and voice recognition company True Knowledge, later called Evi Technologies. Tunstall-Pedoe charts the company as a "10-year adventure", seven as a venture-backed independent start-up before being sold to Amazon in 2012 for a reported £21m and its CEO staying on for over three years. 'When we launched, we were trying to solve the problem of basically using computers by speaking to them or with language,' he recalls. 'When you look at science fiction I watched as a child, Star Trek or Blake's 7, all the computers you just have a chat to. It understands you and is the most natural interface. This is how we interact with people. 'But when you used a computer back then, you select from a menu, click a button or you guess keywords and browse links that come back from the keyword search. So we were tackling the problem of how to use language to interact with machines, which of course is incredibly topical now because large language models (LLMs) have essentially cracked that. But we were doing this way back.' In the mid 2000s, Tunstall-Pedoe would demo his software on conversational search and recalls being told by Google (GOOG) that not only was this a worse customer experience but that keyword search was infinitely better. 'That's definitely not the case now,' says Tunstall-Pedoe, who lives in London and Cambridge, the latter a 'crazy converted church' where Spitting Image was also created. Read More: 'Why we set up a sustainable mobile operator to save people money' Evi pivoted several times as a start-up, from a search engine that answered questions to producing a voice assistant, launching at the same time as Apple's Siri. The 30-strong Cambridge outfit had multiple acquisition offers before working on a secret project for Amazon, now known as Alexa, which went to market in late 2014. He later spent a period as an angel investor backing over 100 tech start-ups before his latest venture, Unlikely AI, came into focus. The British venture, which raised nearly £15m following a seed round in 2023, is currently developing technology that combines LLMs with symbolic methods to make AI safer for companies in sectors such as healthcare to provide verifiable insights. 'How do you enable systems built with modern AI to be fully explainable, auditable, not get you into trouble with regulators, and not lead you to making business decisions that cost you lots of money?' says Tunstall-Pedoe. 'We're tackling some of the biggest problems in AI, which is how to make it completely trustworthy and how to ensure the answer is always accurate.' With the advent of ChatGPT, the inventor has admitted that most companies aren't succeeding when it comes to implementing LLMs, which learns to predict the next word, into workflows. He says: 'If you're making a business decision based on the result of an LLM and your private data, and it gives you a wrong answer, that can be really expensive financially and to your brand. If you're in a regulated industry it can get you fined. You can even get the CEO in jail in extreme cases if you break regulation really badly and this is a fundamental problem.' Read More: We sold a hand cream every 36 seconds after appearing on This Morning Thus Unlikely AI, he adds, was born to not only earn trust but to solve complex issues with its deep learning software. 'That means making it accurate, so when it gives you a result, you can trust it,' says Tunstall-Pedoe. 'Sometimes it may tell you it doesn't know rather than guess and it also means explainability and auditability, to fully explain in a very clear way how the result was created. 'Ultimately, it's also about the customer or business being in control and that's the kind of experience that we're shooting for. We're still relatively early, but we're looking to be very successful and become a very big company.' How I identified the most boring day I built this huge database of millions of facts about the world, where the system could make sense and reason with them. I thought it would be interesting to do an analysis of all the facts that the system knew, and in particular with the goal of finding what the least interesting date in history was — 11 April 1954. The story went so viral that it got told and retold and it's had new life ever since. There was a play in Germany that was kind of themed on it and all sorts of crazy things that have happened as a result of that. Entrepreneurship One of the things I love is that you can have a very significant impact as an individual if you've got an idea and you pursue it. If you build a software product that does something valuable and different, it can scale like no other product can. You can have 100 million users within a couple of months if you create a really successful piece of software, which is kind of almost impossible in almost any other industry. That's what's exciting about it, that potential for big impact if you're developing technology in the software world that really hits home. Read more: The boss who has found 'nature's answer to plastic' Meet the company that finds 'must-haves' to make everyday life easier Impossibrew CEO says Dragons' Den failure sparked alcohol-free brand's riseError in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Yahoo
14-06-2025
- Business
- Yahoo
Meet the Briton who helped teach Alexa how to talk
British AI entrepreneur William Tunstall-Pedoe has always been 'motivated by impact'. He started pushing the boundaries on what's possible with machines as a 13-year-old schoolboy, while Tunstall-Pedoe secretly helped create and launch Amazon's (AMZN) Alexa over a decade ago. In between, his anagram technology was used by author Dan Brown for the Da Vinci Code and he even identified the most boring date in history. The first keyboard seeds were sown at Dundee High School in Scotland where he was able to venture to the technical college next door before classes and during breaks to write computer software on its mainframe. There were financial rewards, too. His computer science teacher, Michael Ryan, had a software business which sold by mail order. However there was no taking advantage of the pupils, says Tunstall-Pedoe, with very generous royalties and a decent income on offer. Read More: Meet Britain's 'king of billboards' who sold his business for £1bn 'I kind of very arrogantly set out to try and solve this problem of using language to control machines,' he tells Yahoo Finance UK, 'to automatically answer questions, change the user experience you have when you interact with search engines, when you interact with computer software, and kind of invented some technology to tackle that.' It is 20 years since he started his search and voice recognition company True Knowledge, later called Evi Technologies. Tunstall-Pedoe charts the company as a "10-year adventure", seven as a venture-backed independent start-up before being sold to Amazon in 2012 for a reported £21m and its CEO staying on for over three years. 'When we launched, we were trying to solve the problem of basically using computers by speaking to them or with language,' he recalls. 'When you look at science fiction I watched as a child, Star Trek or Blake's 7, all the computers you just have a chat to. It understands you and is the most natural interface. This is how we interact with people. 'But when you used a computer back then, you select from a menu, click a button or you guess keywords and browse links that come back from the keyword search. So we were tackling the problem of how to use language to interact with machines, which of course is incredibly topical now because large language models (LLMs) have essentially cracked that. But we were doing this way back.' In the mid 2000s, Tunstall-Pedoe would demo his software on conversational search and recalls being told by Google (GOOG) that not only was this a worse customer experience but that keyword search was infinitely better. 'That's definitely not the case now,' says Tunstall-Pedoe, who lives in London and Cambridge, the latter a 'crazy converted church' where Spitting Image was also created. Read More: 'Why we set up a sustainable mobile operator to save people money' Evi pivoted several times as a start-up, from a search engine that answered questions to producing a voice assistant, launching at the same time as Apple's Siri. The 30-strong Cambridge outfit had multiple acquisition offers before working on a secret project for Amazon, now known as Alexa, which went to market in late 2014. He later spent a period as an angel investor backing over 100 tech start-ups before his latest venture, Unlikely AI, came into focus. The British venture, which raised nearly £15m following a seed round in 2023, is currently developing technology that combines LLMs with symbolic methods to make AI safer for companies in sectors such as healthcare to provide verifiable insights. 'How do you enable systems built with modern AI to be fully explainable, auditable, not get you into trouble with regulators, and not lead you to making business decisions that cost you lots of money?' says Tunstall-Pedoe. 'We're tackling some of the biggest problems in AI, which is how to make it completely trustworthy and how to ensure the answer is always accurate.' With the advent of ChatGPT, the inventor has admitted that most companies aren't succeeding when it comes to implementing LLMs, which learns to predict the next word, into workflows. He says: 'If you're making a business decision based on the result of an LLM and your private data, and it gives you a wrong answer, that can be really expensive financially and to your brand. If you're in a regulated industry it can get you fined. You can even get the CEO in jail in extreme cases if you break regulation really badly and this is a fundamental problem.' Read More: We sold a hand cream every 36 seconds after appearing on This Morning Thus Unlikely AI, he adds, was born to not only earn trust but to solve complex issues with its deep learning software. 'That means making it accurate, so when it gives you a result, you can trust it,' says Tunstall-Pedoe. 'Sometimes it may tell you it doesn't know rather than guess and it also means explainability and auditability, to fully explain in a very clear way how the result was created. 'Ultimately, it's also about the customer or business being in control and that's the kind of experience that we're shooting for. We're still relatively early, but we're looking to be very successful and become a very big company.' How I identified the most boring day I built this huge database of millions of facts about the world, where the system could make sense and reason with them. I thought it would be interesting to do an analysis of all the facts that the system knew, and in particular with the goal of finding what the least interesting date in history was — 11 April 1954. The story went so viral that it got told and retold and it's had new life ever since. There was a play in Germany that was kind of themed on it and all sorts of crazy things that have happened as a result of that. Entrepreneurship One of the things I love is that you can have a very significant impact as an individual if you've got an idea and you pursue it. If you build a software product that does something valuable and different, it can scale like no other product can. You can have 100 million users within a couple of months if you create a really successful piece of software, which is kind of almost impossible in almost any other industry. That's what's exciting about it, that potential for big impact if you're developing technology in the software world that really hits home. Read more: The boss who has found 'nature's answer to plastic' Meet the company that finds 'must-haves' to make everyday life easier Impossibrew CEO says Dragons' Den failure sparked alcohol-free brand's riseSign in to access your portfolio
