A sober look at Artificial Intelligence in journalism from the GEN Summit
Possible challenges and necessary precautions for worst case scenarios
Wednesday’s presentations had an optimistic and engaged attitude to new technology and tools, but to provide another perspective we asked the two moderators to imagine the worst case scenario and offer advice or perspectives on several challenging aspects of the growth of AI.
Talks such as “Mixed reality, machine learning and AI for journalists” moderated by Adam Thomas, Director, European Journalism Centre, followed this pattern with live examples of a new highly algorithm-heavy searchable video library, Video Indexer by Leanne Van Oostende, Manager, Media Solution Sales EMEA, Microsoft and Mixed Reality (MR) goggles by Exozet.
In “It’s raining bots: Four best practices to make the most of automation” moderated by Noriko Takiguchi, Managing Editor, robonews.net, the workings and spread in reach behind El País’ Facebook chatbot were explained by David Alandete, Managing Editor, and Robert Unsworth, VP Americas, News Republic, through a comprehensive list of online organisational engagement techniques highlighted that AIs “are not tools of replacement, but tools of empowerment.”
At the bottom of the article we have listed our previous coverage of the exciting aspects of AI, but now we will explore some of the potential pitfalls:
Perhaps the major fear about AI is that it will force journalists out of a job. Is that a well-founded fear?
(Adam Thomas): There is no job security guarantee, but we have to remember that current revenue problems are not based on AI. You can look at AI from two perspectives, one that AI will replace journalists through cost-cutting due to advertising disappearing, or AI can do much of the basic tasks of the newsroom leaving the tasks that people are good at.
(Noriko Takiguchi): We don’t have to worry about it right now, maybe in another five or ten years. There are still a lot more work that humans can do, that machines can’t replicate. What’s happening in sports, financial and weather reporting, which can be automated is really opening up the schedules for journalists to work on creative thing.
Will there come a time when there will exist two tiers of newsrooms; those who can afford the new standard of tools and those who cannot? Will this be harmful in a similar way as the imbalance between local and global news is now?
(AT): It will be a spectrum as it always is, some have reliable streams of funding and some don’t. Tech will come down in price, but once you decide to invest, make sure it’s a central part of your strategy or vital for your audience.
What gets expensive is when newsrooms pointlessly experiment, if you are going to move into VR it’s potentially expensive, but if you’re ESPN it makes sense, you have an audience and subject matter that demands it.
(NT): Some of the more well-connected organisations can afford to experiment or to deploy products that won’t become accessible to smaller newsrooms until five or ten years later.
One way to work around this is to connect and collaborate with big tech organisations such as Facebook or Google. Hopefully more, smaller organisations will soon be able to create their own experiments or products so they don’t have to rely on that, when they become more knowledgeable and there will be more free tools.
The big technology platforms and tools have creators or investors that might have biases we as journalists disagree with. Are we giving away too much power to them?
(AT): Facebook, for example, is at one end of the extremes as a media company and on the other an unregulated editor of the internet. This creates another spectrum where on the one end you have issues of freedom of speech and on the other a complete lack of transparency and a lot of power moving away from the news organisations.
(NT): We have not thought through this yet, because we need technology and users. We try to shake hands with big tech platforms, which might not have any knowledge about or share our conscious about what journalism should be, so we have to ask ourselves if we can we share all our data with them. Maybe NPOs or some open source organisations might develop the same things, we have to do this, but we have to be careful.
If something can be exploited, someone will exploit it. How can we prepare for the increasing exploitation of the tools as they become more dominant in the industry?
(AT): If your goal is to sow misinformation the opportunities will be there, but on the flip side the USP of a news organisation is trust. The problem we face now is news organisations not leveraging that trust which makes them special, we have not been good at promoting that.
Proper media could use all the same technology available to troll farms to push their content too, but they if they can shine more transparency and focus on trust and get really smart using content, platforms and trust there is no reason they can’t win that battle.
(NT): We already saw this in the American election with bots being used to produce and spread fake news; that it is easy for people to use wrong information, and that subtle opinions can be changed to sway readers in certain directions
But this happens with every technology, some use it in the right way, some use it in the wrong way. We have internet banking which is very useful, but some people lose all their savings when someone exploits the system to steal because it isn’t a perfect system.
What safety advice would you give organisations and outlets for preparing for or surviving in the future?
(AT): What role does trust play for you, are you a trust first organisation, do you bake in a layer of trust into every aspect of your organisation? It’s about authenticity, the process may be correct, but does trust follow through? Think about how you can be transparent about your finances, about your processes and your relationship with your audience.
The second thing is to employ people who understand both network theory (how people use the internet, communicate), and computational thinking (how can I as a human formulate this problem so it can be understood by a computer).
If you can build that understanding and critical thinking into your news organisation then you have a lot of the tools necessary to survive an avalanche of misinformation.
(NT): Prepare by focusing on security of course, you now have to have a security person in the team, someone who is an expert on data protection, confidentiality and so on. Make sure to also get an ethics person who can help uphold strict journalistic standards.
For survival I think the ecosystem will change, maybe outlets will come to alliances, moving from ten papers covering all the same stories to splitting the coverage into areas where every one of them has a specific strength and collaboration on a better whole complete picture.
And the individuals? Will we all have to become basic pay freelancers and retrain every six months to be prepared for the constant changes in technology and fads?
(AT): I see more freelancers than I used to. It strikes me smart people who want to work in news under the age of 35 are going freelance or moving into PR and communications. This brain drain is a real worry.
It is perfectly possible now with a mail chimp account, a medium blog and a twitter handle to set up something that will rival the New York Times technology section or debate or insight section
I think that has to be cherished though and news organisations have to do more to find the 17-year old in their bedroom getting 18 million views on their gaming channel. Don’t just copy this person, employ this person.
It is of course great then that we have these events to pull people together.
(NT): I myself am a freelancer and I have noticed that when the shift to online came, the outlets focused on making sure their sites looked like they were flooded with news, they have eternal amounts of news to create now and we become like factory workers.
The notion that “AI is coming to take our job” is everywhere, but when you look at it closely it is a slow transition, we always have to be coordinating and adapting to that transition. Instead of thinking that AI as a concept is taking our jobs, we have to realise it’s a patchwork of several AIs, and it is augmentation rather than replacement.
Before we fear, we have to understand the components. We need people to teach us what these components are which make this kind of conference very useful.
Which aspect have you picked up on that makes you personally worry for the future?
(AT): The shift away from advertising revenue, news organisations are still not nervous enough about this. A lot of time because the power is stacked at the top of the news organisation, making it difficult to make deep institutional changes that allow you to be more agile.
It’s easy to hire a bunch of AI programmers and say “go do some cool stuff”. It’s hard to” say maybe this business needs to be for the few not the many”, so we can be really focused. We need to orient around delivering to a specific audience and listening to them.
Another big worry is local news; they don’t have the resources or training to really weather the storm. When you lose local journalism you lose community cohesion, political awareness on a really important level. Those things have deep, deep impacts beyond journalism and into the fabric of society.
(NT): I am not very worried, but readers have to train themselves to learn and train themselves, not necessarily to understand the code itself, but to understand the platforms and how they work and how that can be exploited. We need a generally higher level of media, social and technological literacy for everyone.
We must not forget that AI is not just a journalistic challenge
Taking a much more comprehensive look at AI, Christoph Thun-Hohenstein, General Director, MAK — Austrian Museum of Applied Arts, questioned how we as a societal whole should consider our relationship with technology.
In his talk, Hello Robot: A new prohuman agenda for journalism, he reminded the audience that they were constantly surrounded and outnumbered by robots already. Easily by a factor of 1,5 or 2 when we recognise that every mobile phone and laptop is a robot. Even Vienna itself, as a smart city, is a robot we live in.
This topic was explored to its full extent at the first social event of the summit at the MAK exhibition, “Hello, Robot. Design between Human and Machine”. Featuring several stages of art installations it explores levels of past, current and possible future human/machine interaction, relationships and integration.
From Thun-Hohenstein’s we have two types of AI at the moment, automated and specialised, which have their strengths and weaknesses, but as a society we are not taking the prospect of Artificial General Intelligence seriously enough. It will arrive in less than 50 years and be a complete game changer on every level:
“At some point we will reach superintelligence and we don’t have a proper discussion anywhere beyond job automation, not in politics, business or science.
There are two concepts major concepts to consider when we reach superintelligence, Dataism and Transhumanism. Dataism believes that all of humanity can be boiled down to algorithms , the more we understand the brain the more of our own programming and behaviour we understand. At some point computers will be able to emulate, reproduce or replace everything a human is or could be.
Transhumanism on the other hand believes that the only way to stay relevant to the machine is to become the machine, for example using Musk’s brain interface system.”
The main challenge that journalism will be some of the few who can be able to help tackle will be to ensure humanity wakes up to this truth, and start to seriously consider our unavoidable new man-machine hybrid society:
“To avoid falling into any of the potential future traps about AI you have be very investigative, explain pros and cons in a clear fashion, be sure what people expect and they know what to expect and of course deliver the three necessary values of good journalism, i.e. a story that:
1 educates me
2 inspires me
3 gives me a perspective
(maybe also entertains me)”
According to Thun-Hohenstein stronger and stronger AI is unavoidable, but a constructive embrace of the technology is not impossible, it simply demands a better global understanding of its implications:
“We get slices of innovation, but we have no idea where this leads us, we miss the bigger combined influences that reach and change the whole civilisation.
We all have to understand the whole panorama of exponential progress and collectively choose a border between man and machine we feel comfortable with. Maybe that will be when we are 80% machine parts, but if we don’t have this big discussion, this current digital experiment could end badly instead of evolving into a direction that would be sustainable for all humanity.”
*Preceding the summit we have had a look at promising future Artificial Intelligence (AI) in articles such as “Artificial intelligence can’t solve every problem in the media, but it can take care of these”, “From tech to ethics: Will AI be a threat to journalism?” and “The slow and steady rise of AI for news”.