Over the 18th and 19th April the Microsoft team (Phil, Lene and Ashley) joined digital leaders in the BIMA community to investigate the opportunities available through AI and Machine Learning, set firmly in a context of strong ethics and best practice. There was plenty of lively debate over the two days and Phil, Lene and Ashley felt that they got a much deeper understanding of the BIMA community’s needs and aspirations, and some plans were set for moving forward.
The Microsoft team got to grips with broader cultural issues of why reduced sugar in Irn Bru is such a controversial topic and yellow shirts in Glasgow is the uniform of Innovation (move over Tour de France). But we delved deeper too. Here is a brief round-up of the topics discussed:
AI opportunity: the appetite for clients to engage with AI was discussed which covered issues from an understanding of how AI can help, to wanting to be doing the latest thing to making sure the client understands the preparation in using AI effectively (it is a tool that needs to be used correctly and appropriately). There was a general consensus that there was interest from the client community but there is currently not enough knowledge and understanding for a clear vision of how they want to use it. The disappointment in the deliverables of personalisation fed into the sense of apprehension.
Data: this was discussed on many levels from ensuring your data is clean and accurate from the outset to how you use it. And of course the fear factor of making sure your data is secure and used appropriately. There was discussion around the correct volume of data to use and how too much data can lead to finding patterns that are not there. Start small. That Machine Learning is built on probabilistic processing means it will not necessarily be accurate. Not to mention the small problem of data bias! How BIMA and Microsoft can help the community access the right data expertise through upskilling or collaboration was also brought into topic and added to the list of future considerations for our members.
GDPR: We can’t talk about data without GDPR being a pillar of the conversation and it resonated through the two days. There are challenges because people don’t know the data they have already. It was noted that there is a dangerous amount of misinformation around GDPR leading to bad practice in attempts to be GDPR compliant or head in the sand approaches. Notwithstanding, there was a strong consensus that GDPR should be embraced as a brilliant opportunity for people to cleanse their data, its time to say goodbye to mass untargeted mailings. GDPR is going to create more opportunities and better profit margins.
Ethics: naturally the conversation led to Cambridge Analytica and there were comments such as being surprised people were surprised – couldn’t we all see something like this coming? Others noted memories can be short and it will all blow over but there will always be evil overlords to watch out for. It was also raised that the ethics of looking after personal data is not a new thing. Another view was that data is in such mess we will be fine as the clean up job is vast. Conversely, we can see how tech can enable vast data breaches but lets us it to protect our data.
Empathy: the man and machine debate went in different directions. Phil noted in Glasgow that it was the first time he had been in a meeting where someone had called for “machine empathy”: we forgive human error but give a machine very short shrift. We discussed empathy in terms of the end use and creative process.
End use: we need to be transparent about how we use data and talk about it in an open and accessible way. As individuals do we opt for empathy of the many or the few? Can I set my satnav so that I take the route with least environmental impact vs “just get me there fastest”.
Creative process: In Edinburgh the debate kept returning to not losing sight of human learning too. The example was the copywriter who creates 300 versions of the marketing emailer header vs AI doing it for you. It was argued the AI route could save time but those 300 copy lines are an essential learning journey for the young copywriter. The comment “we need to keep the human emotion in it” resonated across the room. The sweet spot is the perfect collaboration of human & tech: human innovation delivered through technical expertise.
The blockers: The conversation turned to what is blocking the way forward to maximising the benefits of Machine Learning and AI - the following issues were shared:
- The quality of data
- Access to, and understanding of, the right tools
- Clients’ understanding of the opportunities available
- Design challenges on presenting data
- Properly understanding what can be achieved through Machine Learning and AI and applying it effectively.
Next steps: Phil, Lene and Ashley suggested coming back organising a hack or two through BIMA to either run through specific product development requirements or work collaboratively on a piece. Addressing a challenge collaboratively that is for the good of Scotland was warmly received, as was looking to collaborate with Scottish initiatives already in place such as Product Forge, the Data Lab and the Machine Learning club.
Thank you to Phil, Lene and Ashley of Microsoft for joining us up in Scotland, and huge thank you to the digital leaders from the following organisations who gave up their time to contribute to the insightful discussion over the two days:
Hot Tin Roof
Software Advisory Services
The Data Lab
This is Millk
Contacting Microsoft and finding out more
E-mail for agencies to get in touch: firstname.lastname@example.org
For all things partner and Data & AI for digital agencies visit: aka.ms/digitalagencies
Events for agencies to come along to to learn more:
1. Azure Technical Briefing – Serverless, Microservices and Containers
Date & Time: April 26, 09:30 – 16:30
Location: Microsoft Edinburgh, Waverley Gate, 2-4 Waterloo Place, EH1 3EG Edinburgh UK
Registration link: https://aka.ms/ukocpisv260418
2. Northern Roadshow events (workshops on Mixed Reality, Data, Business Intelligence, Analytics, Cognitive Services and Bots + interactive Bootcamp session)
- May 2 in Liverpool
- May 3 in Manchester
3. AI Hack May 15-17 (London)
- where you can bring your application/PoC /project and work on it with Microsoft
- If interested, apply here and let the Microsoft team know so they can help with application getting through