Illustrations for my book: AI or not to AI?

Black man and White woman in AI drawn picture

Some of you might know I have been writing a book about my dating experiences.

Its moving along thanks to some great friends who have done such a great job editing, structuring and shaping the book. But one thing I turned my attention to a while ago, is the illustrations.

I did pay for an artist out of my own money but wasn’t quite happy with every single illustration for each chapter, so only had about half done. The rest I’m talking to another artist about but recently been quite impressed with the AI art generators like DALL-E 2, Midjourney, Nightcafe.

The generated works are strange and abstract enough to fit with what I’m looking for in the book. Not only that, the ownership and copyright seems to be working out (from what I read using DALL-E 2).

(c) Copyright. OpenAI will not assert copyright over Content generated by the API for you or your end users.

I certainly seen the AI bias in some of the images generated. For example if I don’t say what gender or race the person is, the AI defaults to male and white. Its only when I deliberately say Black male / female it then switches. I would also say the images of black women are not as fully thought out as white women. Because I’m generating pictures of dating, it always defaults to straight dating unless I add something to the query. Likewise the women are always thin never curvy unless specified. Actually a few times, I got women who were pregnant. Of course every single time I make a query, it takes credit (money) making it costly to really test its bias, sure someones already on this.

The big question I have is, if I was to use DALL-E for illustrations in my book, what would that say or mean for my stance around AI, bias and data use? To be honest, I’m actually thinking about generating the front & back covers in full colour, rather than the in book illustrations.

Maybe I should be less worried about this? Or even better I was thinking about ways to not just make clear it’s AI generated but show the process of selection or something similar?

Thoughts?

Loving my first Tech Open Air Berlin

https://www.flickr.com/photos/cubicgarden/35867186991

I was invited to talk at Tech Open Air Berlin (TOA Berlin) a while ago and the week of the event came around much quicker than I thought it would. Because of plans a while before, I had planned to be London to pick up my Estonian e-residency card, go to a semi-internal BBC AI event, visit “into the unknown” at the Barbican and host a lecture about Databox in R&D London (phew!). It became clear it was better to fly from London Gatwick to Berlin and fly back to Manchester (I couldn’t work out how long it would take to get to Stanstead and Ryan Air worked out more expensive once I factored in luggage). This did meaning 3 hotels over 6 days but it was acceptable in my head.

When I finally flew over to Berlin, the storms delayed the flight and when I finally made it to the hotel I was exhausted but noticed that I hadn’t sent my presentation to TOA or at least my updated version. So spent quite a bit of time checking my email to make sure.

I was on at 10:10am in the main conference and was pretty tired by the time I made my way to the venue which was way in the east and required a few changes from Rosenthaler Platz to Funkhaus Berlin Nalepastrasse. On arrival I was taken to the buzzing speakers lounge where I met Laura, who helped sort things out with some serious help from the tech support guy.

Human & AI Powered Creativity in Storytelling

A look at how Human and AI-powered creativity can be combined to build better storytelling

I felt the talk went ok, but it wasn’t my best because I took too long giving only room for 2 questions from the moderator. I certainly felt if it was a hour or so later it would have been far better. Regardless, it was captured and should be on the TOA Berlin youtube channel soon.

After the talk I was locked in conversation with 3-5 people about the data ethic considerations of adaptive media and how on earth this can work. All fascinating conversations which had to get cut a little short as I signed up to do a Ask Me Anything

TOA Berlin

This took part in a plastic dome within a busy room below the speakers lounge. I wasn’t expecting anyone to show up but there were 2 people waiting for me. I was asked about my role in the BBC and some of research we are conducting. Then a 3rd person dropped in. He said he had read my blog and suddenly there was a moment of “uh oh!” But it was fine, although we talk about data ethics and dating. I’ll be honest the AMA was fascinating and quite refreshing.

TOA Berlin

After this and a lack of lunch (my own fault, talking to people), finally started going to different sessions. Most were rammed and I remember going to Why Supermarkets Must be Replaced, Creators & Audiences: An Open Relationship, Motivating Behavior When Attention Is The World’s Reserve Currency, How far can VR go to enhance your sex life? Future of Sex Podcast with BaDoinkVR and The Future of Collective Governance and From Trump To Universal Basic Income: Leveraging Technology To Understand What Europe Thinks.

TOA Berlin

So quite an amazing cross section of talks and sessions!

I can’t emphasize enough how big some of the spaces are and the whole place just felt like it was buzzing. The engagement was high and everywhere I went people were getting involved. I don’t know the total amount of people, but it felt like a few thousands at least. I can only describe it as what I imagine South by South west is like but in Berlin. Theres so much happening and besides the conference there is a whole music track, expo, corporate spaces and even a thing called open circle only for speakers and vip’s. Its pretty overwhelming, but in a good way.

One of the other fascinating things about TOA was the amount of Fridge events or as they call it Satellite events. You can also apply to do a satellite using a online form and a video chat with TOA.

I attended two and had to miss a few because I needed to move hotel on Friday. Here are the two I attended.

Propellor | Forum #1: Using Tech to create the future of Film

I first met Erwin at the Documentary and Factual World Congress in Sweden late last year when he told me about the Propellor film tech hub.  We kept in touch and when he mentioned the Forum would be during TOA Berlin, I was happy to say I would be there too. I agreed to help by encouraging people to think about adaptive media in a workshop of ideas.

Propellor | Forum #1

The event was a satellite to TOA and was hosted at Price Waterhouse Coopers building only minutes from Berlin Hauptbahnhof (if unlike me you go out the main entrance and not the back exit). It started with networking and some canopies then an introduction followed by 3 5min pitches by myself (Adaptive Media), Jannis Funk (distribution of AV content) and Aljoscha Burchardt (Curation of AV content).

Once the pitches were out the way there was time for a few questions. Most of the questions I got were asking how on earth is adaptive media possible, I answered in a quick 1min breakdown of object based media.

TOA Berlin

After this, people grouped around the pitches they were interested in and the DO school took control.

TOA Berlin

It was good workshop with enough push to get things done in time but also allowing things to emerge from the grassroots. At the end of the workshops ideas were presented to the pitch group and the best was then presented to everyone at the forum.

TOA Berlin

I found the Friemily film a great idea so very fitting for adaptive media.

All the results have been written up at propellorfilmtech.com.

TOA Berlin Satellite: Machine Learning, Trust and Public Service

Machine learning, trust and public service

Myself and Ahmed from the BBC Blueroom put in a proposal around a idea/concept of the public service internet and machine learning. Only a few days earlier the Blueroom had put on a AI & society conference titled BBC Blue Room presents Artificial Intelligence & Society. So fresh from that and some ideas from myself and others in R&D, we proposed the question; where does public service fit in the age of machine learning and the business models which come along with automation and algorithms.

Machine learning, trust and public service

TOA provide the space and we provided the workshop/talk under their brand. The space was an amazing co-working space called Mindspace and they were really helpful with everything. The only down side was the workshop was meant to be more participatory but the room set up didn’t quite work for this. The central chairs were not move-able at all. Ideally more of a circle would have been better really.

Machine learning, trust and public service

We were not expecting a huge turn out but thats exactly what we got. As me and Ahmed got started we focused on the business models which are most associated with machine learning. To be honest we spoke maybe too much and by the time the discussion got going, we ran over time. This is also where the layout of the room became a disadvantage. Regardless I drew up thoughts on a piece of paper and mapped out connected ideas (mindmapped it) while people talked and discussed.

Machine learning, trust and public service

There was a lot of discussion about public service and the point of public service. We got talking about why people choose to work in the public service oppose to commercial companies.

The discussion about trust thew up a whole lot of discussion about fake news and disinformation; someone suggested maybe a trusted public entity could have a trusted index score for sources? Something like snoops which are currently having their own problems. This lead nicely on to the transparency question and the fact public organizations should be more transparent than other companies. Which people felt could mean public organisations could benefit from the transparency in choices and algorithms. I earlier called it xray mode in the conference talk.

Collaboration came up again and again, there was a discussion suggesting we should work with not just ARD/ZDF but also Fraunhofer and others like Arte. There was also a feeling, some pioneers in this sector could share insight and new models with less forward thinking public organisations, who are struggling to keep up with the internet age.

TOA Berlin

As a whole I was very impressed with TOA, its really a festival of tech, art and ideas. I saw solar panels made into art, a stage made in the wood and far too much free redbull.

I would highly recommend TOA to others, think SXSW with a strong European backbone, I will make my way back next year.

Alien intelligence like plant intelligence?

Future Everything 2016

I never got around to writing about the Future Everything conference which is a shame because it was another good conference with plenty of interesting topics and conversation. I really should share my mindmap which is full of interesting thoughts and ideas I picked up while listening to the various sessions.

In the intelligence section Darius Kazemi talked about the bots he creates and how they deliberately don’t have human characteristics. He then raised the question of what is intelligence which is always fascinating (I could spend a whole post just about that alone) but he then pleaded that we should stop trying to humanise them, referring to them as alien intelleigence.

When we are building artificial intelligences, whether they’re corporations or recurrent neural networks, we are building alien intelligences.

There was a bunch of good points like how can we programme them to be human if we don’t really know ourselves? There was also a really good discussion about the ethics, responsibility and diversity of the creator and what is created. This was explored much further by Lydia Nicholas and her work into ethical frameworks for data use.

But I found it interesting to read Matt Locke’s post pretty much saying a similar thing. AI like plants?

…I’m here to talk about a network of conversations that we can’t hear. The garden around us — blossoming fruit trees, thick borders, and fresh cut lawns — is also communicating, an ecosystem sharing information and competing for resources using a grammar and vocabulary that is completely alien to us. Wright thinks we can learn from the way plants talk to design better networks of bots — the intelligent agents that are being hyped as the way we’ll communicate with our tech ecosystems in the future. Instead of building bots like Apple’s Siri and Amazon’s Alexa in our likeness, he believes the answer might be to stop trying to make bots behave like humans altogether.

It’s also interesting the parallels between Darius’s comment about not really knowing whats going on inside the complex neuronetworks we are generating and Matt talking with Tim about the science of plants communicate and it wasn’t till recently we could understand how this actually worked.

Both are worth listening and reading, then consider the parallels.