Artificial Intelligence | News | Insights | AiThority
[bsfp-cryptocurrency style=”widget-18″ align=”marquee” columns=”6″ coins=”selected” coins-count=”6″ coins-selected=”BTC,ETH,XRP,LTC,EOS,ADA,XLM,NEO,LTC,EOS,XEM,DASH,USDT,BNB,QTUM,XVG,ONT,ZEC,STEEM” currency=”USD” title=”Cryptocurrency Widget” show_title=”0″ icon=”” scheme=”light” bs-show-desktop=”1″ bs-show-tablet=”1″ bs-show-phone=”1″ custom-css-class=”” custom-id=”” css=”.vc_custom_1523079266073{margin-bottom: 0px !important;padding-top: 0px !important;padding-bottom: 0px !important;}”]

Learning to Trust AI in Troubled Times

As budgets tighten amidst a global crisis, marketers are scrambling to find better ‘sources of truth’. Whether it’s good prospecting performance, campaign management, or audience optimisation, there are many areas of success when it comes to the programmatic landscape. To meet this need, programmatic advertising is increasingly being driven by machine learning. So why would anyone doubt machine learning?

Machine learning models are, in many ways ‘expert liars’. Machine learning optimises by any means necessary, and if blurring the truth or taking into account irrelevant information helps to optimise, then this is what occurs. It’s scary to think how much an unchecked model could get away with in the fast-paced world of programmatic, where seconds count.

In fact, Artificial Intelligence (AI) researcher, Sandra Wachter, actually calls machine learning algorithms “black boxes”  saying: “There is a lack of transparency around machine learning decisions and they’re not necessarily justified or legitimate”.

Is the truth a matter of interpretation?

So, how can anyone ensure a machine learning model is telling the truth? The best way is to treat the model like a job interview candidate; that is, any statements made should be treated with the due amount of scepticism, and facts must always be checked.

When it comes to performance, everyone wants it better. However, while a model might offer better performance on face value, it’s important to ask how exactly is that measured:

  • Was it measured using test data, simulations, perhaps an artificial trial?
  • Was there any bias?
  • How much budget was involved?
  • What was the optimal KPI and how did others measure?

Machine learning technology can be time consuming and expensive, and it’s remarkably easy to waste money on a bad algorithm. Having good, solid proof a model works is a great way to avoid wasting budget. Fact-checking and asking for more evidence is vital if unsure of results, and if the model vendor can’t offer access to an analyst who can back up the numbers with the work, move on.

Data relevancy and privacy – not just words

Just because all data is accessible, doesn’t mean it should be used, or that each point of data is as important as another.

Related Posts
1 of 1,095

Is knowing whether someone has bought a product before as important as the colour of their socks? If all data in the machine learning model is being used, marketers must ask how and why. Why is all of the data used? Why is all this important? What tests were run to prove it? Is the model even allowed to use all the data?

Everyone’s familiar with the concept and purpose of GDPR and similar global legislation. So, you must make sure you ask the question about how data is being used, or run the risk of severe fines.

Better optimisation for business success

Brands have clear metrics to hit and it’s the job of client services, together with data engineering, to ensure the machine learning optimises towards the KPIs. However, the beauty of machine learning is it frees up the client services team to do more than just achieve the brand’s KPI; it can help brands achieve business goals, too.

With thousands of successful campaigns under their belts, client services know what works and what doesn’t. Users should expect to be able to contact a specialist at any time to make sure it’s doing what the clients want.

The most dangerous thing to hear when discussing machine learning? Silence

When talking about purchasing machine learning with a vendor who can’t (or won’t) answer your questions, it’s time to bail. Marketers must feel empowered to ask any and all questions of vendors, and just like a job interview, if the answer isn’t a good fit then neither is the candidate.

Not knowing about or not understanding machine learning is accepted. However, what’s not acceptable is to not be allowed to question “machine learning just does it”. In order to innovate, especially in volatile environments, everyone needs to better understand machine learning and to achieve this, a two-way conversation is vital.

About Silver Bullet

Silverbullet is the new breed of data-smart marketing services, designed to empower businesses to achieve through a unique hybrid of data services, insight-informed content and programmatic. Our blend of artificial intelligence and human experience …

More about Silver Bullet:

Comments are closed, but trackbacks and pingbacks are open.