Cédric Villani, French mathematician and politician, said in March 2018: “Artificial intelligence will be everywhere, like electricity.” (Source: Nouvel Observateur) Indeed, AI technologies are progressing rapidly, with countries like the USA and China investing massive amounts to develop the field. “Until recently, AI has proven itself by surpassing humans in terms of its analytical capacity and its ability to build relationships and bring out the “message” hiding behind a cluster of complex information. Data analysis and the determination of usage trends on social media are part of this trend […], says Alexandre Beaulieu, PhD Candidate. In the future, I believe that it’s the application of “generative” algorithms – those who, beyond analysis, specialize in creating new content – open the most possibilities. Chatbots, like those who assist the customer service, are the simplest expression. It’s not foolish to imagine a machine producing promotional material specific to each customer, who develops the argument most likely to join him, and who chose the best platform to do it. Or, I imagine that we will see more and more internal social networks of large companies to grow the knowledge accumulated by all resources over the years. So we could think of a robot that determines what problematic an employee is working, and sends solutions that exploit the work done elsewhere in the company that could have passed under the radar of this employee.”
Regarding social media most specifically, the applications of AI are already multiple: choosing the right clothes, communication, data analysis, chatbot, market research, etc. Advanced Symbolics Inc. (ASI), a firm based in Ottawa, is a pioneer of market research and analysis, opinion research and real-time surveys using AI.
Advanced Symbolics Inc. “is the only company in the world that can make represented samples of online populations to accurately predict human behaviour.” (Source: Advanced Symbolics Inc.) Polly, their artificial intelligence persona, has accurately projected the Brexit, the election of Donald Trump in 2016, and the recent mid-terms, and the last Ontarian election results. “Advanced Symbolics scours social media mentions—it looks at aggregate data without storing any personal information—and gathers huge amounts of intel, cross-referencing it against news stories, academic journals, websites, Statistics Canada and whatever else is relevant to come up with an answer to a company’s question.” (Source: CPA) A strong asset of the company is the fact that they don’t collect personal information nor anything that can be linked back to anyone in particular. There is such a critical factor with privacy and personal data in social media nowadays, that the trust that individuals and companies can rely on can be scarce. “First, I don’t advocate for personal, individual AIs that are reading our social media feeds and providing us with ongoing commentary, says Kenton White, PhD and Chief Scientist at Advanced Symbolics. At the best it’s an annoyance, like a spell checker that keeps suggesting the wrong word and I constantly must override. At worst, a personalize AI can be scary, unjustly flagging me or a loved one as “at risk” (for suicide) when they aren’t. […] At Advanced Symbolics, we work with samples of the population. This prevents individuals from gaming the system — they may or may not even be in the sample! Sampling also ensures we are learning from everyone without overfitting to anyone. People come and go within the sample and no one person is ever weighted too heavily.”
The company was hired by the Public Health Agency of Canada for a very important pilot project to research the warning signs of suicide, before tragedy occurred. The three months project was to predict areas with high risk of suicide, by analyzing public posts on social media. The data collected is not tied to an individual, therefore could not be used to stop someone from committing the act, but would rather “be used to inform mental health resource planning.” (Source: Macleans.ca). This is not the first time that AI is used for suicide prevention. For example, Facebook is also working on this issue, as stated in this December 2017 case study.
Lessons for Others
Artificial intelligence is a field that is growing exponentially. The applications of AI in social media, as in the majority of our daily lives, are endless. The work on AI cannot be isolated from a good communication practice. It’s imperative that companies and governments on all levels continue to inform the population on the development of AI, its usage, risks and benefits. We are all impacted by the technology. To this idea, Advanced Symbolics is leading the way by using (of course) their social media platforms to link, retweet and share all kind of information related to their work and to AI in general. Their Twitter feed is particularly interesting, as well as Polly’s blog. The use of humour is also appreciated to lighten up the content that could be dry.
“Eventually, as we learn more, I can see AIs helping governments create better services for their citizens. Imagine a bus system that reroutes buses in real time, ensuring there is a seat for everyone. Or an AI that efficiently allocates beds for homeless people during cold winter nights. What if we could increase security on a subway train just before violence would break out. AI + social could lead to more efficient use of limited resources!”, concludes Kenton White.
Everything that I read to prepare for this case study points out to the same consensus: the use of artificial intelligence in social media is allowing a more efficiently way of using resources: human, material and time.
Advanced Symbolics Inc.
Industry: Artificial Intelligence
Name of Organization Contact: Kenton White, Chief Scientist
Authored by: Cami Beaulieu
If you have concerns as to the accuracy of anything posted on this site, please send your concerns to Peter Carr, Program Director, Social Media for Business Performance.
Wiese, Marie (May 2, 2018) The Grim Future of social media https://www.forbes.com/sites/forbesagencycouncil/2018/05/02/the-grim-future-of-social-media/#25a4b1497b9a
Schaefer, Mark (2018) 10 Ideas driving the future of social media marketing https://businessesgrow.com/2018/03/05/future-of-social-media-marketing/
Borzykowksi, Bryan (October 30, 2018) If you could read my mind https://www.cpacanada.ca/en/news/pivot-magazine/2018-10-30-erin-kelly-advanced-symbolics
Mulvey, James (April 10, 2018) Social Media in 2020: 11 Data-Backed Predictions https://blog.hootsuite.com/social-media-2020/
Smith, Albert (April 13, 2018) Why the Future of Social Media will Depend on Artificial Intelligence https://www.smartdatacollective.com/future-social-media-depend-artificial-intelligence/
Fay, Sophia and Nora, Dominique (February 28, 2018) Cédric Villani: L’intelligence artificielle sera partout, comme l’électricité. https://www.nouvelobs.com/economie/20180228.OBS2848/cedric-villani-l-intelligence-artificielle-sera-partout-comme-l-electricite.html
Anderssen, Erin (November 24, 2018) Can an algorithm stop suicides by spotting the signs of despair? https://www.theglobeandmail.com/canada/article-can-an-algorithm-stop-suicides-by-spotting-the-signs-of-despair/