Picture this: You are crossing the road and hear the buzz of a WhatsApp message. You know you shouldn’t - but take a cheeky glance anyway until a voice booms out: danger - don’t check your phone as you cross the road.
This may sound like the stuff of science fiction or a passage from George Orwell’s dystopian novel 1984 but it’s a reality.
Sensors that see, smell, hear and shout commands have been developed by Brazil-based technology company Polsec. Using artificial intelligence, the system is currently being piloted in the Indian cities of Mumbai, Delhi and Agra.
Depending on your point of view, this is either another worrying move towards total surveillance or deeply reassuring to a country’s citizens.
For Polsec chief executive Renato Werner, it’s definitely the latter.
"People say it's not good for privacy but the sensors don't listen to the contents of conversation - only words," Mr Werner, 45, told The National at the Future Cities Show in Dubai on Tuesday.
The sensor’s cameras can identify people and report on traffic levels. The bionic-ear can detect words such as “help” or “police” and sound the alarm. It can also pinpoint an individual’s voice, gender, ethnicity and even accents. The smell sensors can monitor pollution.
“If you spread these in the city, then you can know everything that is going on," he said.
At the exhibition in Dubai, Mr Werner pitched his system to a country embracing AI and smart technology. Cameras would appear on practically every building and along roads.
Only a few weeks ago, Dubai Police announced proposals dubbed “police without policemen” which could see a blimp, CCTV cameras and rapid response teams replacing officers on the street, so Polsec's system could be just around the corner.
The sensors are all contained in one unit and hundreds, if not thousands, would be spread across a city or country. The hardware for one costs about US$200 dollars. The system could be used to cut traffic, reduce crime and even prevent a chemical attack.
The India pilot began last October. While concerns have been raised in India about the nature of the project and the results have yet to be announced, Mr Werner and Polsec are confident.
“We will not reduce the crime rate - we will take it to zero. If a human knows that if he does something wrong and he will be caught, then they won’t do it," he said.
More than 25 years ago, Mr Werner was working in his own video camera shop in Brazil. Forget AI, this was the VHS era and before the mass proliferation of the internet.
One day in 1995 a police officer came in and requested a spy camera he’d seen in a James Bond film. The camera was worth two years' revenue alone. Mr Werner travelled to Japan, came back with one and sold it to the officer. He soon realised the appeal of the defence and security industry so he switched and Polsec followed.
In 2011, the company’s work using camera technology lead to a 70 per cent crime reduction in Manaus within just six months.
“It got results. It’s not surveillance for its own sake," he said.
He points to Europe as a place where his technology could be implemented, and closer to home. With an aging population in places, if someone fell on the street, AI sensors could trigger call for help.
“This is a chance for Dubai to ensure no crimes go unpunished and it can easily claim to be the smartest safe city, even as it continues to expand and its population increases.”
Across the world there is a growing concern over the use of surveillance and the dangers that unrestricted use of AI might bring. There is also a worry about these types of systems falling into the wrong hands. According to the Polsec chief, the algorithms cannot be used in the wrong way because it follows local rules and it doesn’t record video or conversations. Instead it snaps one frame per second and listens only for keywords.
“AI is like water – it’s essential. But that same water can also be used to torture. But we can write an algorithm that says - don’t do that.
“If you have good intentions, follow the rules and are not a criminal, the technology is there to help. It doesn’t scare me.”
Tesla founder Elon Musk recently said that AI is something we should be afraid of. But Mr Werner rejects this and urges strong international regulation to address these concerns – right up the level of the United Nations.
“I respect him and my next car will be a Tesla but I disagree. Humans are the creators of AI, it’s a code - regulation stops it becoming the monster. If that happens, then you go to the jail.”
He also said that operators cannot access the system and these privileges can only be granted by judges.
“Today you are already being tracked. You can say it’s right, I can say it’s wrong, but we should follow the rules. If it’s regulated and implemented for the good, then everyone will want it.”