Pong Roadtrip to EGG London 2019
Last week I went on a Pong Roadtrip to EGG London 2019 (#EGGLDN2019 powered by Dataiku) with four other colleagues from Pong. This event is as Dataiku itself calls it a; “Human-centered AI Conference”. Why ‘Human-centered’? Because no matter what, humans play an important role when it comes to artificial intelligence. That sounds strange when you assume that the machine will take over ‘everything’. However, as Florian Douetteau (CEO of Dataiku) pointed out in his keynote, there is an explicit responsibility for the human when it comes to AI. An explicit responsibility to deliver meaningful AI applications.
By definition, it’s not about how great, state-of-the-art, an organization’s infrastructure is. The amount of data is not important either. A mature AI solution adds real value to the organisation. This organization actually uses the application in question. That ultimately determines the success of AI.
Why EGG London 2019?
The EGG event is a worldwide travelling circus. In London we could expect the following from EGGLDN2019; 3 themes around:
- Ethical Challenges And Achieving Responsible Enterprise AI
- Empowering People And Organizations To Lead A Business Transformation
- How They Actually Do It: Real-World Use Cases
“Something full of life and potential, that stands for birth (or rebirth) and change”
EGG is not just an abbreviation. EGG stands for; “…something full of life and potential, that stands for birth (or rebirth) and change.”. Successfully implementing AI within an organization is not something that will happen overnight. It’s a whole process. Data Science is not magic. It’s not a question of making cool things. AI must really add value.
With the EGG London 2019, Dataiku has created a very diverse event. It varied from keynotes, very diverse presentations and round table sessions, during which various specialists gave their views on all kinds of themes around Artificial Intelligence. For a software supplier there were surprisingly few tool sessions. Of course, two hands-on sessions could not be missed. Based on a case and Dataiku DSS, a Machine Learning process was developed in two parts.
The role of Artificial Intelligence
For a software supplier, Dataiku takes a rather critical look at their own role in AI. How does AI stay fun and how do you prevent AI software from being used for the wrong things? Software has to signal and report when algorithms develop negatively and are potentially going to be used for the wrong things.
“Change only happens with Purpose”
The fact that AI is going to be used with the wrong intentions is one of the risks that is recognised. Other risks manifest themselves in the area of privacy and ethics. Biases in algorithms must also be eliminated. How do you tackle these risks? Make sure you do a thorough analysis and evaluation of your organisation. Know what you want to do with AI. Caroline Carruthers had an interesting quote: “Change only happens with Purpose”. Why do we do things?
Artificial Intelligence in practice
Datascience is team sport. I am convinced that datascience does not exist as an individual function. Different roles are important in the entire datascience process. Communication and Stakeholder Management are essential. The involvement of the management is also (or perhaps even; specifically) very important. Where necessary, management must be educated and trained to fully understand what the datascience function entails.
Show that AI projects can be successful. Think big, but start small. Show what it yields after each short iteration. Communicate and celebrate the successes. Show what it costs, but certainly also what it delivers. And software? With the help of software you can make yourself redundant. That you only have to solve a puzzle once. With software, you can keep things simple. It saves work and the need to document afterwards. Although, of course, it’s no excuse not to document.
“A person without data is just another person with an opinion”
What makes it so interesting and fun to work in the world of datascience and artificial intelligence? All specialists indicate in their own way that solving puzzles with statistics, software and data is their passion. I understand that. The computer should not be seen as a threat. Because of all the innovations that take place around data, there is a window of opportunities. Opportunities for people to do smart things with data. I don’t know who said it, but: “A person without data is just another person with an opinion” does fit well in this context. But don’t let the algorithms go their own way. They are smart.
Special attention for a presentation by Sandra Wachter of the University of Oxford. She gave a presentation about; “Online advertising and ethics – Affinity Profiling and Discrimination by Association in Online Behavioural Advertising”. It’s definitely worth reading.
In conclusion, it was a very interesting day. Dataiku EGG London 2019 was great. I Learned a lot and gained a lot of inspiration. What came up at each session, is the important role of the human being. As we say at Pong: “IT’s a people’s business!”.
Would you like to talk more about this subject? Coffee or tea is made like this and the fridge is filled. You are most welcome at the Pong office.
Originally published on the Pong-blog.
Thanks for reading.