AI and the media: Q&A with John Barnes, William Reed, UK

In preparation for an in-depth report on AI and the media, we spoke to several experts in the field via email. One such expert is John Barnes, Chief Digital Officer at William Reed.

What are your current most common use cases of AI?

Editorial:

Content aggregation / Related content (identify trending stories and other stories associated with them to give better context)
Cross-posting on multiple properties (a story of BD.com could be automatically posted to NI.com when it becomes relevant to the NI audience increasing the ROI of historical content – long tail)
Content auto-tagging (In the past did a one-shot use of IBM Watson https://www.ibm.com/cloud/watson-natural-language-understanding to auto-tag all the GB news stories and build a cleaner version of the tags dictionary)
Help checking grammar, formatting, outbound links, SEO metrics and recommending relevant images from a DAM, all before putting the news story live
Content translation, podcasts and videos auto-transcription to improve SEO
JWPlayer Cloud, our video publishing and hosting solution provides some fancy AI features as well (article matching, automated video thumbnails – like Netflix does): https://www.jwplayer.com/video-engagement/#article-matching

Marketing:

Audience segmentation (AC, Piano, DMP, etc.) and calculated metrics (Piano give us for example’ likelihood to subscribe’ https://docs.piano.io/lts-likelihood-to-subscribe/, ‘likelihood to return’ https://docs.piano.io/likelihood-to-return/, ‘likelihood to cancel’ https://docs.piano.io/ltc-likelihood-to-cancel/)
Content personalisation & recommendation (TG tested Biblio https://www.biblio.ai/ and we now have Piano Content https://docs.piano.io/piano-content-overview/ and Composer available for the eZ sites using Piano – former CXENSE product)
Chatbots (we did a POC a few years ago but costs of commercial solutions are a bit hard to justify)
Email marketing (NL content personalisation and sending hours personalisation based on when people are reading their emails https://business.adobe.com/products/campaign/ai-powered-emails.html)
Web analytics alerts (anomalies detection) and user-friendly chatbot query interface to ask for metrics or insights. See more about ‘Adobe Sensei AI and machine learning capabilities in Adobe Analytics’ https://experienceleague.adobe.com/docs/analytics-learn/tutorials/data-science/using-machine-learning-and-ai-for-the-moments-that-matter.html?lang=en
Personalise marketing ad campaigns copy/creative based on content context and user audience. There a few vendors proposing this sort of adtech: https://www.adcreative.ai/, https://www.bannerflow.com/
A/B split testing augmented with AI. See capabilities of Adobe target https://business.adobe.com/products/target/automation.html
Being able to engage audiences with customised experiences
Optimise and scaling these experiences based on performance Frameworking documentation

Legal:

Fight copyright infringement by automating the process of monitoring for duplicated content in the media space

How else will you use AI in the coming year?

Our most likely use cases include:

Speech to text
Anomaly detection
Document Summarisation
Pattern/trend analysis
Building on our work with predicting customer behaviour based on audience segments crossed with behavioural attributes and conversion factors.
Automatic tagging of data
Automated tagging of images
AI supported software testing
AI supported data quality analysis

What excites you most about the longer-term future of AI, and why?

Removal of some human weaknesses: Efficiency and accuracy/bias should be completely removed with the right models
Being able to efficiently deliver personalised one-to-one customer experiences at scale

I have thought about the possibility of an artificial intelligence. For a long time.

When editing computing magazines in the late 1980s and early 1990s I wrote a lot about AI and machine learning. Whilst in their infancy, the design, implementation, training, and analysis of outputs were very involved tasks.
It is exciting how much the the field has developed and how some of those first problems have now been addressed – the choice of activation function, the number of processing nodes, number of input and output nodes – these are all pretty much nailed down now. Furthermore, we can snow elect from a wide range of pre-trained models in the commercial arena and treat them as “just another” component in a software stack .

What concerns you most about the longer-term future of AI, and why?

The disappearance of human relationships, austere world. We can already see how people interact with their smartphones. What will happen if we give them the opportunity to interact more with AI ?
The risk on authority that AI-driven content could bring due to an over-reliance on technology as this could be disruptive in all of the wrong ways.
There appears to be a moral and ethical vacuum surrounding AI. There are a wide variety of extremely valid concerns that are not being adequately addressed, by philosophers, ethicists, and specialists in the field of AI. It is incumbent on them to provide thought leadership and stimulate informed public debate – without such debate, politicians have no motivation to stay ahead of the subject area. If politicians are not onboard then legislation, when and if it is required, will be rushed, fudged and inadequate

What are some of the best AI tools (already available for use) that you can recommend to media companies to investigate immediately?

Azure AI services
Python – Tensorflow and sci-kit learn
ChatGPT
Salesforce Einstein
IBM Watson