Presenter: Vishnu SARAN, Co-founder and CEO, Invigilo Technologies
Category
🤖
TechTranscript
00:00I'm vishnu and i'm the ceo and cofounder of invisible ai.
00:05So invisible in latin means look after, and the whole
00:09Premise of the company is to look after workers' safety.
00:12And that's a topic which i'm going to discuss today.
00:15How do we use ai to help improve workplace safety?
00:19And why does this mean a lot to me? i grew up in a family where
00:23My dad is in manufacturing, i have relatives in construction,
00:27And safety has been a common language that we have spoken
00:30From day one, and it's been a big problem that's been bugging
00:34Many countries across the world, and it is something that we
00:38Have to address with or without ai.
00:41So this is a typical construction site, a lot of
00:45High-risk activities, a busy manufacturing plant, warehouse,
00:49And oil rigs, one of the most dangerous locations in the
00:52World. And what is one common thing
00:55Across all of these industries? every single industry has
00:59Shocking statistics of accidents.
01:01One out of four workers in this space suffers a major accident
01:06At least once in a lifetime, and this is crippling to the
01:10Families and the companies. About 2,600 people are dying
01:15Every single day because of this, you know, the dangerous
01:19Nature of these industries. And this affects personal lives,
01:23Affects the people's families, and a lot of other personal
01:26Implications to the workers. On top of this, from a
01:30Financial standpoint, about $250 billion is lost every single
01:35Year because of workplace fatalities.
01:37And why is this happening? you know, we're living in a
01:40Progressive world with a lot of safety protocols and things
01:43Happening that help improve safety standards.
01:47So me and my cofounder, we went down to oil rigs, we stayed
01:51There, we stayed in construction sites, we stayed in deserts,
01:55You know, really high-risk places, to really understand
01:58What is happening, and that's how invisual was born.
02:01And the key problem that these guys face all the time was a
02:05Severe lack of visibility on what's going on because sites
02:08Are just so big, you can't really know what is going on
02:11Everywhere. And secondly, you know, when
02:14There's an inspection going on, workers behave a lot safer.
02:17You know, they automatically become good behavior.
02:20But the moment workers are not around, sorry, the moment
02:23Safety managers are not around, it goes back to, you know, lack
02:27Of following protocols. And on top of this, safety
02:31Enforcement and patrols are very manual, tedious, and time
02:35Consuming. So a bit more of myself,
02:39Graduated from the national university of singapore in
02:422015, spent one year in silicon valley doing startups and
02:46Interning in a few fast-paced companies.
02:49I cofounded with my partner in 2021, and four years into the
02:53Journey, and happy to announce that we've been getting a lot
02:56Of traction and recognition on regional and local media.
03:00So moving on to, like, you know, how ai has actually been used on
03:04The ground by the industry. Firstly, you know, voice to
03:09Text. Then you have text analytics,
03:11And lastly, you have video analytics.
03:13All these are different forms of, you know, ai technologies
03:16That all of us have been interacting with, and even
03:20Industrial companies have been using.
03:23Even simple things like, you know, siri or google or using
03:27Chat gpt, these are the most basic and, you know, daily use
03:31Forms of artificial intelligence. But how can we use the same, you
03:35Know, concepts in industrial work sites?
03:38And that's what i'm going to talk about today in a bit.
03:42So one of the first things that we have done at invisilo is to
03:46Use ai to identify near misses. And this is the concept of using
03:50Video analytics, where we have developed a software that
03:54Analyzes camera streams in real time to detect workplace
03:58Fatalities and hazards. Things like people walking close
04:02To open edges, people walking in the line of fire of moving
04:07Vehicles, missing barricades, missing ppe, and the list goes on.
04:12Every industry has their own, you know, set of rules and
04:16Concerns that could be identified with simple, you
04:20Know, ai technology that could be deployed.
04:23Secondly, it's been heavily used in predicting, you know,
04:27Machinery breakdown, faults, and for predictive maintenance.
04:30A lot of industrial work sites have tons of machines, but
04:33There's no way to monitor the health of the machines.
04:36So by simply deploying sensors, analyzing the sensors using
04:40Machine learning algorithms, a site team now has visibility on
04:45What, you know, to do next. And, you know, we have also
04:49Started using ai to recreate safety lapses.
04:52Generative ai to create algorithms or create safety
04:55Scenarios, mimic the various types of hazards, accidents that
04:58Are not possible in real life. So this is then how we used to
05:02Teach the ai to detect such things in the future, although
05:05Such events have not happened before.
05:07So this is a way of, like, creating fake data sets for the
05:11Algorithms to learn on top of it. So with that, you know, how does
05:16Ai and what does it mean to workplace safety?
05:19It changes more from a reactive space to a more proactive way
05:22Of detecting safety hazards. It centralizes the experience
05:26Of numerous safety officers across the world.
05:29Imagine one central brain of safety managers from thousands
05:33Of years of experience and, you know, various industries across
05:35The world. And it's also a safety officer
05:39That never sleeps and does not make many mistakes.
05:42But at the end of the day, you still need a human to verify
05:45And take the action. It doesn't do the whole
05:48Preventive action for you. It just informs you of what to
05:51Do next. And with that, you know, i'd
05:55Like to say that ai is here to stay.
05:57And being used in the right manner, having a very user
06:00Driven approach, staying with the customer on the ground is
06:04How you understand and really tweak ai to help improve
06:07Workplace safety. And with that, thank you so
06:09Much.