Incentivising social enterprise through technology

Updated: Nov 27, 2019

What do you do?


“I do nothing.” Rudradeb calmly states.


I’m thinking this might be a tough interview, but fortunately my fear of sounding stupid keeps me silent.


“It is my belief, that when we live our lives, it should feel as simple and easy as if we’re doing nothing.”


That was like a philosophical kick in the face. I’m relieved I can now play off my precautionary silence as profound contemplation and agreement.



Rudradeb runs Omdena, a decentralised project management ecosystem to engage volunteers of varying skill, using machine learning to solve real world challenges. More to the point, Omdena only accepts these challenges, or projects, from socially positive enterprises.



First, back to doing nothing,


It sounds coherent enough, that if you do nothing your mind is a tranquil basin ready to absorb new information (listen) and think spontaneously (creativity).

From LinkedIn:



“The idea of keeping us constantly busy means that your mind is full, so it can’t come up with new ideas. When empty, you can be creative so can perform really well.”


He emphasises being conscious of the question: “How crucial is what I’m doing? Is this really necessary?”


You can read more about this ‘doing nothing’ ethos in a recent post of his on LinkedIn, here.






It’s in part why he came up with Omdena, he explains, for he was travelling the world, and the world of AI, and summarises for me his three eureka observations.

Three observations of AI:


1) There’s little AI for good. “The majority of people see AI as a tech that takes away jobs, such as building autonomous robots. There are very few people in the space of doing AI for good, solving social problems and creating real value for society.”


2) Education has changed significantly. “No matter where you come from, thanks to online education now everyone has access to the same knowledge. There’s been huge democratization of knowledge and it has leveled the playing field.”


As a tangent of point two, he mentioned “I also observed that these entities have a lot of knowledge, but lack experience. Companies are still hiring on the basis of CVs. So there is equal knowledge but not equal opportunities” – there are still a premium for Cambridge candidates, but a perfect world is that no matter where you live [or what education you have access to] you have equal opportunities.”


3) Competition inhibits collaboration. “In the work environment, or everywhere, with everything, school, university, work…wherever, ‘competition’ is everywhere. Constant competition doesn’t make us work together, it doesn’t make us collaborate. Like, in machine learning, with competitive platforms like Kaggle, there are so many people in the world who don’t have people to collaborate with...when you make it collaborative you unlock so much more power.”


In fairness to Kaggle, a recent interviewee of mine (Andrey Lukyanenko) told me that it was common for multiple teams to group together at the end of a competition and combine their work for an academic submission. I thought that was awesome; you can read more about how we have gamefied progress here.



So, to summarise.

1. There’s little AI for good.

2. There‘s a democratisation of knowledge (but not opportunity).

3. Collaboration is powerful.


Now’s probably a good time to tell you that Omdena is an 1) ‘AI for good’ ecosystem that takes on only pro-social projects, taking advantage of 2) a democratisation of machine learning knowledge around the world, connecting them through an automated project management tool that facilitates 3) global remote collaboration.


Volunteers use machine learning to solve real social problems in the form of challenges on Omdena.


“Bringing people together from different parts of the world, different time zones… just giving them a problem to solve. It’s significant.”


How do you incentivise the volunteers, I ask.


“There are many people, even at the top of the pyramid, who have intrinsic motivation to do good for the world and learn something. We just bring them together.“


It’s free [smart] labour. He reached his free virtual workforce exclusively through word of mouth on social media.


“I wouldn’t define it as a profit or non-profit. Obviously it needs to be sustainable, it has some costs, so it has to make money. But it shouldn’t be for profit, I would define it as a ‘social enterprise’.”


How do you generate revenue?


“Some of these initiatives, about 8 of our challenges – like the United Nations one for example, they don’t have funding. So we think, ok, how can we make it work.”


“On the other hand, there was a startup identifying trees that could fall on cables and cause fires. So the project was to protect against forest fires. It was a startup so they received funding and this, for example, covers the other free project for the UN.”


So as long as the business overall has positive cash flow, they’ll take on free work for social causes, however those fund bringing businesses must also be pro-social.


“I personally think that all businesses should be driven by social good.”


“I don’t want to build a business around government grants because I don’t think that’s sustainable. We want to build a bottom up model where communities come together and solve their own problems.”


Omdena’s medium channel is entirely community written.


He clarifies he’s not anti-capitalist, “I grew up in India and saw how much capitalism can benefit. The free market is good, but not to monopolistic levels.”


Of the last 10 challenges Omdena took on, 8 were for free.



What’s next?


“The next stage is to scale this up. We are talking with a US senator about building a model to reduce child abuse. We’re also talking to UN. My goal here is every second week we start a new challenge, next can we start 1 every week. Then take it further.”


“Our costs are very minimal”, explaining how the team personally take under market salary.

Below market rate is an attractive proposition (for clients), so I ask how he chooses between them – how he prioritises who’s challenge to accept.


“We base it entirely on the impact of the solution. We also have two conditions:

1) the organisation must have the capacity to implement our solution, and

2) we ask how big the impact is going to be. How big is the problem, and how many people are going to be impacted by the solution?”


Beyond those two filters, the community itself decides which projects to take on.

Omdena is literally an online organisation or business ‘shell’ that structures the community, enabling project management by assigning volunteer members and assigning others to check up on them, and so on.



I ask if this low ratio and cost base is sustainable.


“This is a very new concept, so we had to prove it. Moving forwards I hope to have 50/50 (paid/unpaid). So, every challenge we have, another is free.”



As someone focused on how algorithms are changing human society, I’m fascinated by the decentralised organisational approach of Omdena and suspect Rudradeb is on to an absolute game changer. In my opinion, this is capitalism at it’s best: companies of the future like Omdena provide a huge financial incentive to do social good – in this case free machine learning labour for pro-social enterprises, so this is an example of capitalism correcting itself.



Incredibly, the community even self-selects workers. I’m beginning to think Rudradeb wasn’t being entirely philosophical about ‘doing nothing’ ;)


“We have rolling applications. 200 people currently. 70 people on the waiting list. New applications for upcoming challenges... we already have waiting lists! Sometimes we prioritise taking people from previous challenges, and roll excess applicants to the next one …if we have free slots!”


I’m told, in total, Omdena has had applications from over 3,000 people in over 75 countries. 400+ people from 65 countries have actually contributed to code base that has been deployed to solve various real world challenges.



Are all solutions rolled out, I ask, thinking back to his earlier filter question (‘does this entity have the capacity to deploy the solution?’).


“Not all! We’re only 6 months old. Whenever we finish a challenge, we hand it over.”


He gave an example of a safe city initiative to fight sexual harassment in India. They generated a heat map of the data that resulted in an application to find the most optimally safe route home.

It sounds like this channeling of volunteer coding hours is making a real difference.


What’s been your most interesting or favourite project?


“They all are quite interesting! They all have an impact, all solve a social problem.”

Whatever. All parents have favourites ;)



Omdena’s success goes against everything we’re taught about management in the workplace.


“It’s self organised. I’ve built a few processes here and there but basically it runs itself.”


“Projects are split into tasks and these tasks are split, and if someone isn’t active for a while then we remove them.”


Actually this part is very important, removing non-active volunteers. It’s important to make it feel like a exclusive community.


“We also bring mentors in if people are stuck on something. There’s a group of 20 people or so who conduct peer to peer review processes as the project progresses.”



It’s the future.




Extra reading:


https://medium.com/omdena/building-ai-for-good-by-the-people-for-the-people-d98ad78b5001

https://towardsdatascience.com/why-kaggle-is-not-inclusive-and-how-to-improve-it-6e2013c82b63

111 views0 comments

Recent Posts

See All

Machine learning will leave no stone unturned

  • LinkedIn

View Privacy Policy  © Machine Commons 2018