AI will completely transform society in future, starting with a reinvention of the way we interact with machines
While AI-based humanoid robots are unlikely, we’ll see a growing number of devices that incorporate machine learning, leading to new guidelines on AI ethics
Businesses need to prepare for an AI-powered future that will enable automation and vital data analysis
Artificial Intelligence (AI) will completely transform the way we communicate with devices in future, as we see machine learning integrated into more devices. While we’re unlikely to see a humanoid cyborg that we can talk to like it’s a person anytime soon, interfaces will need to be completely redesigned to incorporate the new technology.
“We’ve got better at building machines that learn from data,” said Jeremy Wyatt, Professor of robotics and AI at the University of Birmingham, at a recent Intel panel discussion on the relationship between AI and humans. “We’re better at building task-specific AI. We already have an ‘ecology of machines’ and each of these will have a little bit of autonomy – using machine learning to do a very specific task”. Comparing the array of AI-powered devices to a rainforest where thousands of different species adapt to their environment and carry out different tasks, Wyatt said: “We’re going to create an AI zoo, not an AI human”.
“We’re going to create an AI zoo, not an AI human”
Moderated by Georgie Barrat, UK TV presenter on Channel 5*’s The Gadget Show*, the panel discussion featured a number of AI and robotics experts, along with co-creator, writer and producer of the UK’s Channel 4* programme Humans, Sam Vincent, who described AI as “an exciting tool that needs to be handled very carefully”. His sci-fi drama explores what it really means to be human, focusing on the wide-ranging implications of introducing humanoid robots called ‘synths’ into society.
“Our human ability to anthropomorphise is incredibly powerful,” said Vincent. While it may be natural to attribute human traits to machines, especially those that appear to show human-like capabilities such as speech, interfaces will need to be completely transformed in future. “We’ve spent 30 or 40 years working out how to interact with computers, and now we’re going to have to reinvent it all again to support AI,” said Wyatt, who added, “When machines are autonomous, they’re not predictable”.
Human-to-machine interactions can be vital in the adoption of new technologies. Wyatt illustrated this with examples of the Tamagotchi* digital pet where the “interaction was incredibly addictive for a lot of people”, to Microsoft*’s annoying and universally loathed ‘Clippy’ cartoon paperclip assistant. Perfecting how humans interact with AI systems will be vital for businesses in future, as the technology becomes incorporated into an ever-increasing number of devices, including autonomous vehicles.
While AI may be the future, there’s some way to go before it becomes accepted in society. Only 9 per cent of people were familiar with the term ‘machine learning’ in research carried out by the Royal Society Machine Learning Working Group, said Sabine Hauert, a member of the group and a researcher and lecturer at Bristol Robotics Laboratory* and the University of Bristol*. The research suggested that people are happy to see the introduction of AI-based systems that can help them but that they have concerns over trust and privacy. “Figuring out what people want and don’t want from the technology is really important,” said Hauert. While doom-laden sci-fi tales such as The Terminator*, along with grave warnings from prominent tech experts raise questions over the potential threat posed by artificial intelligence, do we really need to be afraid of AI?
“Wary is fine, afraid is not,” said Chris Feltham, Industry Technical Specialist at Intel. “We need to have the right controls in place – it’s important to give people information and to ensure their privacy,” he said. “Most people know that there’s some sort of AI involved with Netflix* and Amazon* recommendations, but they probably couldn't tell you how it works. It’s the same for voice assistants like Amazon Alexa – people are just starting to become more aware of this category”.
The experts agree that hands-on experience of new AI-based technology and robots is going to play a key role for the technology’s acceptance. “Experience is going to be the great teacher,” says Wyatt. And while machine learning and AI may not yet be universally understood or accepted, it’s already being used in a number of innovative ways to make people’s lives better. One example is the Center for Missing and Exploited Children which is using Intel-powered AI to help speed up the search for vulnerable children.
The organisation gets around 8 million tips a year, but only has 25 analysts. This means it can take a long time to sift through all of the data it receives, sometimes taking valuable weeks to follow up on a vital and time-sensitive lead. “They’ve applied AI to the problem and brought that time down from 30 days to just a day,” explained Feltham. “And because they’ve got this capacity for other skills like image recognition, they can start bringing up other data sources and training AI to trawl images on websites”.
Change.org, the world’s largest online petition platform, is also using artificial intelligence to fuel positive change in the world. AI algorithms are used to help to identify people that may be interested in a particular issue, based on petitions they’ve interacted with in the past, in order to boost engagement. These examples show how organisations can use AI technology to automate tasks and gather better insight from data.
Intel is supporting the adoption of AI technologies in businesses in a number of ways, including a range of machine learning software for developers and its Intel® Nervana™ AI Academy. This is targeted at developers, academics and start-ups and aims to democratise AI innovation by providing the necessary training and tools.
While it’s important to be prepared for innovation in AI, the transition won’t happen overnight. “A lot of this technology will be phased in gradually,” said Feltham, comparing it to the roll-out of autonomous driving technology. The SAE (Society of Automotive Engineers) International sets out a 0-5 scale detailing each level of autonomous technology where 0 represents humans doing all the driving and 5 refers to total automation. A similar system could perhaps be adopted for the gradual introduction of AI.
“I’d love to see this technology become ubiquitous, as an assistant for humans that everyone can use,” said Feltham. Vincent thinks that AI’s power lies in interpreting big data, saying: “When these AI tools can analyse and interpret data on a huge scale, such as the global economy or climate, and make suggestions on how to manage them, it could be radical”. However, he is somewhat pessimistic as to how humans would react to such advice from a machine. “I can see a scenario where we point an AI at a really complicated human problem and then it gives us a recommendation that might not suit us, and we ignore it”. One way or another, it’s going to be fascinating finding out.
*Other names and brands may be claimed as the property of others
For more information: